Secondary recurrent miscarriage and H-Y immunity

1. Henriette Svarre Nielsen*

+Author Affiliations

1. 1.

The Fertility Clinic 4071, University Hospital Copenhagen, Blegdamsvej 9, Rigshospitalet, DK-2100 Copenhagen , Denmark

*Correspondence address. E-mail: henriette.svarre.nielsen@rh.regionh.dk

Received August 17, 2010. Revision received December 21, 2010. Accepted January 31, 2011.

Next Section

Abstract BACKGROUND Approximately half recurrent miscarriage (RM) cases remain unexplained after standard investigations. Secondary RM (SRM) is, in contrast to primary RM, preceded by a birth, which increases the transfer of fetal cells into the maternal circulation. Mothers of boys are often immunized against male-specific minor histocompatibility (H-Y) antigens, and H-Y immunity can cause graft-versus-host disease after stem-cell transplantation. We proposed the H-Y hypothesis that aberrant H-Y immunity is a causal factor for SRM. METHODS This is a critical review of the H-Y hypothesis based on own publications and papers identified by systematic PubMed and EMBASE searches. RESULTS SRM is more common after the birth of a boy and the subsequent live birth rate is reduced for SRM patients with a firstborn boy. The male:female ratio of children born prior and subsequent to SRM is 1.49 and 0.76 respectively. Maternal carriage of HLA-class II alleles presenting H-Y antigens to immune cells is associated with a reduced live birth rate and increased risk of obstetric complications in surviving pregnancies in SRM patients with a firstborn boy. In early pregnancy, both antibodies against HLA and H-Y antigens are increased in SRM patients compared with controls. Presence of these antibodies in early pregnancy is associated with a lower live birth rate and a low male:female ratio in subsequent live births, respectively. Births of boys are also associated with subsequent obstetric complications in the background population. CONCLUSIONS Epidemiological, immunogenetic and immunological studies support the hypothesis that aberrant maternal H-Y immune responses have a pathogenic role in SRM.
Key words

secondary recurrent miscarriage

H-Y antigen

pregnancy outcome parity

pregnancy complications Previous SectionNext Section

Introduction Recurrent miscarriage (RM) defined as three or more consecutive miscarriages affect 13% of females (Tulppala et al., 1993; Katz and Kuller, 1994). Approximately half of the cases remain unexplained following standard investigation (Quenby and Farquharson, 1993; Stephenson, 1996).

Random chromosome errors such as trisomy, monosomy and polyploidy, are responsible for 50 80% of miscarriages in the general reproductive population (Jacobs et al., 1987; Ohno et al., 1991; Morales et al., 2008) and these random events complicate the research of the etiology of RM. Important for the search of non-random causes of RM are the findings that the frequency of abnormal chromosomes in miscarriages decreases with increasing number of miscarriages (Ogasawara et al., 2000) and the risk for having a chromosomally normal miscarriage is increased after one chromosomally normal miscarriage (Warburton et al., 1987; Stephenson et al., 2002). Distinguishing between primary RM (PRM) and secondary RM (SRM) may reduce the heterogeneity of RM patient populations. Approximately 40% of the women with RM have given birth to a child prior to the series of miscarriages and accordingly they are diagnosed with SRM (Jivraj et al., 2001; Christiansen et al., 2006). During pregnancy fetal cells enter the maternal circulation (Evans et al., 1999; Adams and Nelson, 2004) and in late pregnancy apoptotic syncytiotrophoblast debris is normally shed in large quantities (several grams per day) from the placenta (Huppertz et al., 2002). SRM is hence preceded by a possible priming of the immune system of the mothers that may theoretically lead to harmful immunological reactions against the semi-allogeneic fetus. A well-known example of immunization in an ongoing pregnancy that can cause harm in subsequent pregnancies is the production of maternal antibodies against the rhesus antigens on the fetus red cells causing fetal erythroblastosis. Similarly, it is possible that maternal immunization against male-specific minor histocompatibility (H-Y) antigens carried by a male fetus in a pregnancy that went to the third trimester may harm (in particular male) embryos and fetuses in subsequent pregnancies. Maternal immune recognition of H-Y antigens has been demonstrated following pregnancies with boys (Verdijk et al., 2004; Piper et al., 2007;van Halteren et al., 2009). Anti H-Y immunity is held responsible for the increased risk of graft-versus-host disease (GvHD) in male recipients of stemcell transplantation with female donors (Flowers et al., 1990;Gratwohl et al., 2001). Two placebocontrolled, randomized trials testing intravenous immunoglobulin treatment for RM in consecutive, eligible patients with four or more miscarriages, carried out at our clinic, found 34 (74%) of SRM patients had given birth to a boy prior to the miscarriages (Christiansen et al., 2002) suggesting that birth of a boy predisposes to SRM. On the basis of these findings, we proposed the H-Y hypothesis that aberrant H-Y immunity initiated in a prior long-lasting male fetus pregnancy is a causal factor for SRM. To test our H-Y hypothesis, we initiated a series of relevant studies in 2003 based on the SRM patients from the Danish Recurrent Miscarriage Clinic in collaboration with Dutch and US laboratories specialized in H-Y immunity as well as The National Institute of Public Health, Denmark. This review presents and critically discusses: (i) studies addressing the association between female donor pregnancy history and the risk of GvHD in allogeneic stem-cell recipients and the immunological priming of females against H-Y antigens as a result of pregnancy; (ii) epidemiological, immunogenetic and immunologic studies in patients with SRM and population-based studies testing the H-Y hypothesis; (iii) detailing the hypothesis followed by a discussion of strengths, limitations and perspectives of the findings.
Previous SectionNext Section

Methods A critical review of the current literature forming and testing the H-Y hypothesis in SRM patients was performed. The review is based on own publications and papers from other groups identified by systematic searches of the PubMed (19682010) and EMBASE (19802010) databases and identifying relevant studies published in English. In addition abstracts from ESHRE meetings were checked and reference lists of identified papers. The latest search was done August 2010. The following MeSH terms were used: H-Y antigen, female, sex factors, risk factors, transplantation, GvHD, pregnancy, parity, pregnancy complications, abortion habitual, abruption placentae, birth order and siblings. This review is based mainly on human studies and animal studies when human studies were lacking.
Previous SectionNext Section

H-Y antigens, pregnancy and transplantation

Donor sex and transplantation outcome

Allogenenic hematopoietic stem-cell transplantation has proved to be a curative therapy for patients with hematological malignancies, though associated with high morbidity and mortality. Donor factors affecting morbidity and mortality have been studied in details to improve transplantation outcomes. Female donors are reported to increase the risk of GvHD (Storb et al., 1977; Bross et al., 1984; Randolph et al., 2004) and transplant related mortality (Gratwohl et al., 1998, 2001). Pregnancy-induced alloimmunization was hypothesized as the underlying mechanisms and female donor pregnancy history and GvHD was thus investigated. Table I gives an overview of human studies addressing an association of the pregnancy history of female donors and the risk of acute and/or chronic GvHD. Despite no clear distinction between parity and gravidity, pregnancy history in female donors were shown to increase the risk of acute (Atkinson et al., 1986; Gale et al., 1987; Flowers et al., 1990; Nash et al., 1992) and chronic GvHD (Atkinson et al., 1990; Carlenset al., 1998; Kollman et al., 2001; Remberger et al., 2002; Loren et al., 2006) compared with other sex combinations of recipients and donors, except from one small study (Przepiorka et al., 1999). Atkinson et al. and Kollman et al. showed a dose response association between increased number of pregnancies among female donors and increased risk of acute GvHD and chronic GvHD, respectively (Atkinson et al., 1986; Kollman et al., 2001). The majority of the studies referred to in Table I was performed on HLA-matched siblings so the development of GvHD was triggered by non-HLA differences between donor and recipient. Male-specific minor histocompatibility (H-Y) antigens were accordingly suggested to be such targets.

Table I
View this table: In this window In a new window

Human studies addressing an association between female donor pregnancy history and the risk of GvHD in allogeneic stem cell recipients.
H-Y antigens

Genes on the Y chromosome encode H-Y antigens. Each of these H-Y genes has an X chromosome homolog that is more than 85% identical at the amino acid level (Lahn and Page, 1997). H-Y antigens were described when it was observed that male skin grafts were rejected by syngenic female mice (Eichwald and Silmser, 1955). The H-Y antigen is expressed as early as the 8-cell stage in mouse embryos (Krco and Goldberg, 1976). Generally, the H-Y antigens are ubiquitously expressed in male cells, including fetal and trophoblast cells (Warren et al., 2000).
Clinical relevance of H-Y antigens in transplantation

H-Y-specific CD8+ cytotoxic T lymphocytes (TCTL) were first described in humans more than three decades ago. A strong post-transplant TCTLresponse specific for male donor HLA-matched target cells was found in the peripheral blood lymphocytes of a female patient who rejected the bone marrow of her HLA-identical brother (Goulmy et al., 1976). Subsequently, H-Y-specific TCTL were found to increase during GvHD in sex-mismatched stem-cell transplants (Mutis et al., 1999). In the latter study, donors were also tested and some were observed to have substantial levels of anti-H-Y TCTL. All were female bone marrow donors and the relatively high levels of circulating H-Y-specific TCTL were suggested to be the result of either pregnancies with boys or prior blood transfusions. Also, H-Y-specific CD4+ T cells have been found to be crucial in females rejecting male grafts (Zelenika et al., 1998; Spierings et al., 2003), as well as in sexmismatched GvHD (van Els et al., 1990a; Faberet al., 1995). A potent B cell response against the H-Y antigen DBY has been shown in 50% of male patients who received stem cells from female donors as they developed a high-titer antibody response to the HY-protein DBY (Miklos et al., 2004). Evidence that the immunogenicity of H-Y antigens results in a coordinated response involving B cells and T cells was subsequently provided (Zorn et al., 2004). Finally, the presence of H-Y antibodies to one or more of five recombinant H-Y antigens correlated with chronic GvHD in male patients with female donors (Miklos et al., 2005). Thus, both cellular and humoral H-Y immunity is associated with GvHD in male recipients of female donors.
Immune priming against H-Y antigens in pregnancy

Fetal cells can cross the placenta during pregnancy and male DNA is detectable in the maternal circulation both during and after an ongoing pregnancy and can persist in the circulation up to 27 years post-partum (Bianchi et al., 1996, 2001). Early reports identified H-Y-specific TCTL in women in whom pregnancy was the only possible exposure to H-Y antigens (Singal et al., 1981; Tekolf and Shaw, 1983). Further support for pregnancy as a source of inducing H-Yspecific TCTL has been given recently (Table II). James et al. (2003) detected measurable levels of circulating H-Y-specific TCTL that readily expanded in vitro in a multiparous female donor who had been exposed to H-Y antigens in her three pregnancies with male fetuses compared with a nullipara woman. Verdijk et al. (2004) found functional H-Y-specific TCTL of a memory phenotype in two of six healthy female donors who had given birth to boys 20 and 22 years previously; H-Y microchimerism was also observed in one of these females. Piper et al. (2007) tested 35 female donors and demonstrated functional H-Y-specific TCTL responses in 37% of the women who had given birth to boys and the prevalence increased to 50% in women with two or more prior births of boys. Thus, pregnancies with male fetuses can prime for H-Y-specific immunity. However, not all women with a previous birth of a boy developed cytolytic activity against H-Y antigens (Verdijk et al., 2004; Piper et al., 2007). A recent study focused on whether natural exposure to fetal minor histocompatibility alloantigens from prior male fetus pregnancies induces different T cells in healthy parous female donors. The presence of functionally different types of H-Y-specific CD8+ T cells, i.e. T regulatory cells (TREG) and TCTL was studied (van Halteren et al., 2009). Indeed, H-Y-specific TCTL were identified in 5 of 10 female donors. Functional H-Y-specific TREG were detected in four of the remaining healthy female donors with a previous birth of a boy and these women were classified as tolerant to H-Y in contrast to females with predominantly TCTL were classified as H-Y sensitized (van Halteren et al., 2009). It remains to be established what causes some women to be sensitized instead of tolerant to H-Y antigens. A recent study found maternal immunity against antigens in the seminal fluid but not in semen (Moldenhauer et al., 2009), whether unprotected sexual intercourse can prime anti H-Y responses is unknown.

Table II
View this table: In this window In a new window

Human immunological studies investigating whether pregnancies with male fetuses primes immune responses directed against H-Y antigens.
HLA restriction of H-Y antigen presentation

Isolation of H-Y-specific T cell clones has identified HLA class I and II alleles that restrict the presentation of the epitopes. To date, the following HLA alleles have been reported to functionally present H-Y peptides and will in the following be referred to as H-Y-restricting HLA class I alleles: HLA-A*01, -A*02, -A*33, -B*07, -B*08, -B*52, -B*60 and H-Y-restricting HLA class II alleles: HLA-DRB1*15, -DQB1*0501/2, -DRB3*03 (Hambach et al., 2007).
Previous SectionNext Section

SRM patients Differentiation between PRM and SRM seems relevant, as there are several significant differences between these two subsets of RM. The frequency of abnormal embryonic karyotypes is significantly lower in patients with SRM compared with PRM (Coulam et al., 1996). Immunotherapy with intravenous immunoglobulin has had no demonstrable effect for PRM patients but has improved live birth rates in SRM (Hutton et al., 2007). However, a new randomized placebo-controlled trial on intravenous immunoglobulin in unexplained SRM did not find a significant higher live birth rate in the treated group (Stephenson et al., 2010). This difference in immunotherapy efficacy suggests that immunologic disturbances are more pronounced in SRM, or that immunological disturbances in the two subsets of patients are different. Disturbances in adaptive immunity may play a role in SRM while disturbances in innate immunity may be of importance in PRM (Christiansen et al., 2008). This is supported by a few large studies. SRM patients, in contrast to PRM, carry the immunological high responder allelles HLA-DR1*03 (Kruse et al., 2004) and HLA-DRB1*15 (Takakuwa et al., 2003) more frequently.

These alleles may present trophoblast-derived peptide to maternal autoreactive T cells or the association to SRM may be caused by linkage disequilibrium to alleles in other loci in the HLA region predisposing to hypersecretion of cytokines with embryo-toxic or trophoblast-inhibiting activity (Raghupathy, 1997). Alternatively, the association is a result of HLADRB1*03 in linkage disequilibrium with a 14-base pair sequence polymorphism in the HLA-G gene, which is associated with RM (Hviid and Christiansen, 2005; Kolte et al., 2010).
Previous SectionNext Section

SRM patients and H-Y immunity Table III gives an overview of the studies testing (or relevant for testing) the H-Y hypothesis that aberrant maternal H-Y immunity is a causal factor in SRM. These studies are described and discussed in details in the following sections.

Table III
View this table: In this window In a new window

Studies in SRM patients testing (or relevant for testing) the H-Y hypothesis that aberrant maternal H-Y immunity is a causal factor for SRM.
Epidemiologic characteristics of SRM
Impact of the sex of children born prior to the SRM diagnosis

Two studies on strictly unexplained SRM patients with three or more miscarriages have found SRM more frequently preceded by birth of a boy than a girl, which suggests that firstborn boys or associated factors represent a risk factor for SRM (Christiansen et al., 2004; Nielsen et al., 2008a). One smaller study found that 27 (47%) boys preceded two or more consecutive miscarriages (Weintraub et al., 2005). The latter study included patients not categorized as unexplained and with only two prior losses which may dilute the estimate of risk factors in casecontrol studies (Christiansen et al., 2005). The studies by Christiansen et al. and Nielsen et al. also explored whether the sex of children born prior to the SRM diagnosis impacts the chance of a live birth after SRM. Patients who gave birth to a boy compared with a girl prior to SRM had a significantly reduced chance of a live birth both in the first pregnancy after referral (Nielsen et al., 2008a) and long-term cumulative chance (Christiansen et al., 2004). Thus prior birth of a boy is a risk factor for unexplained SRM and the negative association to the subsequent chance of a live birth suggests a causal relationship.
Sex ratio

The sex ratio (male:female ratio) prior and subsequent to SRM has been investigated in a 20-year cohort of unexplained SRM patients (Nielsen et al., 2010b). The sex ratio among children born prior to SRM was 1.49 compared with 0.76 in subsequent births, P < 0.0001. Both sex ratios differed significantly from the sex ratios in the control populations. The sex ratios were even more skewed in patients expected to have a low incidence of aneuploid conceptions as all their miscarriages were at gestational Week 10 or more. The sex ratio of births prior and subsequent to this subgroup of SRM patients was 2.31 and 0.21, respectively. These data suggest the existence of a male-specific factor triggering SRM and making pregnancies with a male fetus more likely to be miscarried after the first birth. This is at odds with results from several studies which examined the chromosome results of miscarriages from RM patients (not differentiating between SRM and PRM) that found an excess of female miscarriage products (Halder and Fauzdar, 2006; Kano et al., 2009). The excess of female miscarriage products in these studies is most likely the result of maternal contamination, which is supported by a recent large scale study where microsatellite testing were undertaken in all miscarriage samples with 46 XX and the corresponding maternal DNA leaving only true fetal samples. This normalized the sex ratios among the miscarried embryos from patients with RM (no differentiation between PRM and SRM) (Stephenson et al., 2009). To support the belief that

euploid male embryos are at an increased risk of miscarriage compared with female embryos are the results from a large study of the anatomic sex ratio of 662 miscarried singleton embryos and fetuses that found the sex ratio was 1.30 (299 boys:230 girls) among miscarried fetuses with normal anatomy whereas the sex ratio was 0.92 (59 boys:64 girls) among malformed miscarried fetuses (Byrne and Warburton, 1987). Sex ratio as high as observed prior to the series of miscarriages is reported from countries with a strong tradition of preference for sons, for example China (Zhu et al., 2009). Low sex ratios have been observed in populations exposed to severe stress such as severe peri-conceptional life events (Hansen et al., 1999) although not replicated in a recent study in a stable western population (Khashan et al., 2009). Emotional stress imposed by repeated pregnancy losses (Bagshi and Fridman, 1999) may explain the low sex ratio after SRM. We found the sex ratio subsequent to unexplained PRM 1.18 in a recent study on pregnancy outcome according to thrombophilia in RM [(Lund et al., 2010), sex ratio data not shown in article], which speaks against the stress hypothesis explaining the low sex ratio in births subsequent to SRM. The different sex ratio after PRM and SRM suggests different mechanisms behind the two types of RM.
Obstetric characteristics of birth prior and subsequent to the SRM diagnosis

Obstetric details regarding births prior and subsequent to SRM may contribute to the understanding of unexplained SRM. Four studies have reported on obstetric characteristics of the birth preceding the SRM diagnosis. Birth prior to SRM were characterized by lower than expected birthweight (Christiansen et al., 1992), a higher than expected frequency of pre-eclampsia (Weintraub et al., 2005) and fetal death (Yang et al., 2006). The largest study found stillbirth, preeclampsia, placental abruption, severe hemorrhage, birthweight <2500 g and preterm birth more frequent in births prior to unexplained SRM than in firstborn singleton controls (P < 0.04; Nielsen et al., 2010b). Among SRM patients, significantly more births were complicated if the child was a boy compared to a girl (44% versus 31%, P = 0.01). The higher frequency of obstetric complications in male fetus pregnancies preceding the miscarriages was also reflected in the mean weight of boys born prior to SRM which was 244 g lower (P = 0.0001) than the mean birthweight of boys born in the control group while this difference was 78 g for girls (Nielsen et al., 2010b). Thus, a large proportion of births prior to SRM and particularly those of boys are severely complicated. Prior obstetric complications potentially influences subsequent pregnancies as the complications are associated with increased production of inflammatory cytokines, systemically or locally in uterus (Gerber et al., 2005; Girardi et al., 2006;Germain et al., 2007). Fetal antigens in the maternal circulations are thus to be presented to the maternal immune system under inflammatory conditions possible causing maternal sensitization instead of tolerance towards the paternally inherited antigens (Steinman et al., 2003). Obstetric complications in births after SRM have been described in two studies. Jivraj et al. observed preterm birth, perinatal mortality and small for gestational age to be more frequent among 67 children born after the SRM diagnosis although only the latter reached statistical significance (Jivraj et al., 2001). The other study found increased frequencies of hypoxia, preterm delivery and birthweight <2500 g among 213 births after unexplained SRM compared with second-born singleton controls (Nielsen et al., 2010b). In contrast to pregnancies before SRM, births of girls after SRM were more often complicated than births of boys, 25% versus 13%, P = 0.03, which is also reflected in the mean birthweight of girls born after the miscarriages being 235 g less (P = 0.0001) compared with the control group. The RM population represents a population at high risk of obstetric problems with a need for close surveillance and girls are more likely to survive albeit with a higher frequency of complications.
Immunogenetic characteristics of patients with SRM
Impact of H-Y-restricting HLA

H-Y peptides can be presented to CD8+ TCTL by HLA class I (present in almost all cells) resulting in direct destruction of the cell to which it is bound. Alternatively, H-Y peptides are presented to CD4+ T helper lymphocytes by HLA class II on antigen presenting cells resulting in the T cell secreting lymphokines followed by both antibody formation and activation of TCTL. Three studies

have explored the hypothesis of aberrant maternal H-Y immunity as an underlying mechanism in SRM and recurrent placental abruptions based on patient carriage of HLA class I and II alleles known to present H-Y antigens (Nielsen et al., 2007, 2009, 2010b). The first study identified eight patients who had experienced recurrent severe placental abruptions in an 18-year national cohort of patients with SRM or repeated second-trimester losses (Nielsen et al., 2007). The patients had a total of 22 placental abruptions in 18 of which the fetus died. Fifteen (68%) of the placental abruptions involved male fetuses. Seven of the patients had a firstborn boy. The frequency of HY-restricting HLA class II alleles among patients were compared with healthy parous controls. Haplotypes with H-Y-restricting class II alleles comprised 64% of the HLA haplotypes in the seven patients compared with only 28% among 37 controls with no pregnancy losses (P = 0.009). Thus, carriage of H-Y-restricting HLA class II is associated with the rare and distressing condition of recurrent severe placental abruption in addition to recurrent pregnancy losses. The second study associated carriage of H-Y-restricting HLA to the prospective chance of a live birth after referral for unexplained SRM (Nielsen et al., 2009). Maternal carriage of H-Yrestricting HLA class II alleles significantly reduced the chance of a live birth in SRM patients with a firstborn boy [odds ratio (OR): 0.17 (0.10.4), P = 0.001] compared with those with a firstborn girl. Among patients with a boy prior to the miscarriages the chance of a live birth was reduced in a doseresponse manner; thus, maternal carriage of one H-Y-restricting HLA class II alleles reduced the chance of a live birth [OR: 0.46 (0.20.9), P = 0.02], while carriage of two alleles resulted in an even further reduction [OR: 0.21 (0.10.7), P = 0.02] compared with those with no H-Y-restricting HLA class II alleles. Live birth rate was not different according to the sex of the child born prior to SRM in patients without H-Y-restricting HLA class II alleles. No reduction in chance of a live birth was found when limiting the analysis to the cases where the firstborn child but not the mother carried a H-Y-restricting HLA class II allele (Nielsen et al., 2009). The last study tested maternal carriage of H-Y-restricting HLA class II and obstetric complications in 213 births after unexplained SRM (Nielsen et al., 2010b). The mean birthweight was 381 g lower (P = 0.006), the gestation 0.9 week shorter (P = 0.06), and the risk of stillbirth, pre-eclampsia and placental abruption increased (P = 0.05) in SRM patients with H-Y-restricting HLA class II alleles and a boy rather than a girl before the miscarriages. No difference in frequency of obstetric complications were found according to sex of the first child in patients without H-Y-restricting HLA class II alleles (Nielsen et al., 2010b). Thus, maternal carriage of H-Y-restricting HLA class II is associated with a reduced chance of a subsequent live birth in a doseresponse manner among SRM patients with firstborn boys. If a subsequent birth is obtained after SRM maternal carriage of H-Y-restricting HLA class II is associated with obstetric complications in patients who, prior to the miscarriages, gave birth to a boy. These results indicate that an aberrant maternal immune reaction against fetal H-Y antigens plays a role in SRM. Of note is the observation that birth of a boy prior to SRM only seems to impact future pregnancy outcome if the patient carries the H-Y-restricting HLA class II as no difference in outcome was observed according to sex of first child in patients without these alleles. HLA class II but not class I H-Y-restricting alleles impact the pregnancy prognosis which may reflect the participation of CD4+ T cells providing help for the CD8+ cytotoxic T-cells in their response against H-Y antigens (Fig. 1). Presence of CD4+ T cells with anti-recipient activity rather than CD8+ T cells was earlier reported to increase the risk of GvHD or graft rejection (van Els et al., 1990b; Zelenika et al., 1998; Spierings et al., 2003). Future studies may identify more HLA class I and II alleles, that may be able to present the various H-Y antigens. Those included in the analysis cannot be considered exhaustive and identification of other H-Y-restricting HLA alleles may alter the results. However, it is possible that the number of H-Y-restricting HLA class II alleles is limited and that the currently identified are the dominant ones (van Els et al., 1992; Hambachet al., 2007).


Figure 1

View larger version: In this page In a new window

Download as PowerPoint Slide

A possible scenario at the feto-maternal interface in an HLA-DRB1*15 positive woman with previous RM due to harmful anti-HY immunity. The figure depicts the possible afferent phase of immunization where the DDX3Y protein from male fetal cells and trophoblast debris is engulfed and processed to peptides by maternal macrophages. Their surface HLA-DRB1*15 molecules present the peptide to maternal CD4+ lymphocytes, which provide help for maternal CD8+ and B-cells. These induce through cellular and humoral mechanisms (that might not be HY antigen-specific) cell death in the fetus that induced the afferent reaction or in subsequent fetuses. Mat M, maternal macrophage; Mat Th cell, maternal T helper cell.
Immunological characteristics of SRM patients
HLA antibodies

Maternal recognition of fetal (paternal) antigens reflected by the presence of HLA antibodies in maternal blood are found in approximately one-third of normal successful pregnancies (Ahrons, 1971; Balasch et al., 1981;Regan et al., 1991) while patients with PRM are found to have a prevalence of HLA antibodies of maximum 10% (Beard et al., 1983; Power et al., 1983;Johnson et al., 1984). This difference in prevalence has been taken as an indirect proof of the hypothesis that a failure to produce HLA antibodies was part of the underlying cause of PRM (Beard et al., 1983). SRM patients on the other hand may have high prevalence of HLA antibodies (McIntyreet al., 1984). Recent studies by Steinborn et al. found an association between an increased prevalence of HLA antibodies and both gestational diabetes and placental abruption, suggesting increased humoral immune response of the mother against the fetus as part of the pathogenesis of the conditions (Steinborn et al., 2004, 2006). A recent study investigated HLAantibodies in patients with SRM and controls (Nielsen et al., 2010c). HLA class I and/or class II antibody responses were significantly more frequent in SRM patients with a boy prior to the series of miscarriages (62%) compared with SRM patients with a firstborn girl (29%, P = 0.03) and compared with PRM (23%, P = 0.02) and healthy female controls (25%, P = 0.005). Among SRM patients HLA-antibodies were significantly more frequent if the births prior to SRM were obstetrically complicated compared with those who had uncomplicated births prior to SRM. Of the pregnant RM patients who were HLA antibody positive in early pregnancy, 41% had a live birth compared with 76% of those with no HLA antibodies. Adjusting for the number of prior miscarriages the chance of a live birth in RM patients with HLA antibodies was reduced compared with patients without these antibodies [OR = 0.22 (0.070.68), P = 0.008]. This study showed a remarkable higher frequency of HLA antibody positive SRM patients with firstborn boys compared with other RM patients and parous controls. This emphasizes that subdivision of RM patients is important not only into PRM or SRM patients but also according to sex of the child born prior to SRM. This study is the first showing an association with HLA-antibodies in early pregnancy and a reduced chance of a live birth in patients with RM. The mechanisms behind this association are unknown. It remains to be investigated whether HLA antibodies are the direct cause of the increased miscarriage frequency or whether it is an epiphenomenon that reflects a

series of immunological disturbances possible based on prior abnormal transfers of fetal cells to the maternal immune circulation. Increased microchimerism as a consequence of previous pregnancy complications (Lo et al., 1999; Leung et al., 2001; Zhong et al., 2001;Khosrotehrani et al., 2003) may play a pathogenic role.
H-Y antibodies

The presence of IgG antibodies to one or more of five recombinant H-Y antigens has been shown to correlate with chronic GvHD in male recipients of stem cells from female donors (Miklos et al., 2005). A recent study investigated H-Y antibodies in SRM patients and controls (Nielsen et al., 2010d). Figure 2 is a heatmap visualizing the H-Y antibody response in patients and controls using a color code correlating to optical density (OD) measures of the ELISA analysis. The frequencies of H-Y-specific antibody positive were significantly higher in SRM patients: 39 (46%) compared with female controls: 7 (19%, P = 0.004), and PRM patients: one (8%, P = 0.01). H-Y-specific antibodies were detected in 33 (49%) of SRM patients with a boy and in 6 (38%) of those women who delivered a girl prior to the miscarriages (P = 0.33). The influence of H-Y antibodies in early pregnancy was analyzed in the 77 RM patients who were in the early stages of pregnancy at the time of serum sample; 43 (56%) of these pregnancies ended with a live birth. Live birth rates were not significantly different in H-Y antibody positive patients (48%) compared with H-Y antibody negative patients (61%, P = 0.26). Only two (12%) of the children delivered by H-Y antibody positive patients were boys, which is significantly lower than the 12 (44%) boys delivered by H-Y antibody negative patients (P = 0.03) and the 51% boys among newborns (Khashanet al., 2009) (P = 0.002). The frequency of H-Y-specific IgG antibodies was significantly increased in SRM compared with both control females who previously had given birth to boys and PRM patients. The presence of H-Y antibodies in early pregnancy was associated with a low sex ratio at birth but not a statistically significantly increased clinical miscarriage rate in RM patients. These results suggest a direct, early (preclinical) and male-specific embryotoxic response in H-Y antibody positive RM patients. The impact of H-Y antisera has been shown in studies aiming at noninvasive techniques for sex selection of preimplantation cattle embryos to increase profitability of dairy and beef cattle production (Ramalho et al., 2004). Between 80% and 87% of murine and bovine male embryos that are cultured in high-titer rat H-Y antisera, at the morula stage, stop their development in contrast to female embryos (Utsumi et al., 1993; Ramalhoet al., 2004). Larger studies are needed to confirm the findings of this first inventory and it also remains to be investigated whether the presence of H-Y antibodies in early pregnancy or around conception of women with no history of RM correlates with the sex of the fetus.

Figure 2

View larger version: In this page In a new window

Download as PowerPoint Slide

H-Y and H-X antibody responses in serum samples from patients with unexplained recurrent miscarriage and healthy women. Heatmaps visualizing the antibody response in OD units to each H-Y and H-X protein of each participating individual, grouped according to patient or control status and sorted with the highest mean OD at the top. Positivity is defined as OD

0.1. H-Y-specific responses are responses directed at the H-Y protein and not the corresponding H-X protein. For patients who were pregnant at blood sampling, pregnancy outcome is given: G, girl; B, boy; 0, miscarriage. For control women pregnancy history is given: 2B, Given birth to only two boys; 3B, Given birth to only three boys; NP, never pregnant.
Previous SectionNext Section

Sex of prior children and subsequent reproductive performance in the background population
The hypothesis of aberrant maternal immune responses against H-Y antigens as a cause of pregnancy related problems was challenged in population-covering studies exploring the association of birth of boys and subsequent obstetric complications. Population-covering studies are possible in countries like Denmark with National Birth and Discharge registries based on unique ID numbers of every citizen. Such studies cover the background population and are advantageous as these large data sets allow testing for rare conditions and small effect sizes. There is no clear biological differentiation between miscarriage and stillbirth. On the basis of the Danish Birth Registry, it was tested whether delivery of boys increased the risk of a subsequent stillbirth between 558 314 second to fifth-born children of whom 1952 were stillborn. The risk of stillbirth increased by 12% after deliveries of boys compared with girls, relative risk = 1.12 (95% CI 1.02 1.23) (Nielsen et al., 2010a). Births of boys are thus associated with both subsequent miscarriage and stillbirth. Also based on the Danish Birth Registry, differences in birthweight of 545 839 second to fourth-born children were noted in relation to sex of older siblings. One or two preceding boys, respectively, reduced the mean birthweight of later-born boys by 29 g (P = 0.0001) and 38 g (P = 0.0001) and later-born girls by 17 g (P = 0.0001) and 21 g (P = 0.0001) compared with later-born siblings with no preceding boys (Nielsen et al., 2008b). Similar findings have been reported from the Norwegian Birth Registry, where it was an unexpected finding in a study with a different aim (Magnus et al., 1985). These differences are 1020 times less than the differences in children born by H-Y-restricting HLA class II positive SRM patients with a firstborn boy compared with a girl. Combining data from the Danish and Swedish Birth Registries regarding second-born children only, preterm birth was more common in second borns with an older brother compared with an older sister, hazard ratio = 1.10 (1.07, 1.13) <0.0001. The results were similar in both data sets, the use of two large population-covering registries lower the possibility of selection bias, and focusing only on second borns reduces confounding (Mortensen et al., 2011). In all three studies the associations did not appear to be confounded by maternal age, interpregnancy interval, or by maternal characteristics that do not vary from one pregnancy to the next. In line with these data, two recent studies have found that having either older brothers or a twin brother reduces the lifetime reproductive success of their siblings (Lummaa et al., 2007; Rickard et al., 2007). Prior birth of boys is associated with a decrease in birthweight, an increased risk of stillbirth and preterm birth, and reduced reproductive success among subsequently born children in the background population. These findings support the hypothesis of non-tolerated maternal immune responses against H-Y-antigens as a possible mechanism behind SRM and also some so far unexplained obstetric complications in the background population.
Previous SectionNext Section

The H-Y hypothesis in SRM

Detailing the H-Y hypothesis

On the basis of the knowledge obtained reviewing the evidence for the H-Y hypothesis, we propose the following pathophysiologic mechanisms for the maternal non-acceptance of the fetal allograft in the SRM patient due to anti H-Y immunity. During the patients first ongoing pregnancies fetal cells enter the maternal circulation (Evans et al., 1999; Adams and Nelson, 2004) and in late pregnancy apoptotic syncytiotrophoblast debris is normally shed in large

quantities (several grams per day) from the placenta. After being processed by maternal dendritic cells, peptides derived from fetal antigens, e.g. HLA or H-Y are presented in local lymph nodes to CD4+ and CD8+ T cells as recently suggested (Adams et al., 2007). In normal pregnancies, this presentation takes place under non-inflammatory conditions resulting in T lymphocytes becoming tolerant to fetally-derived peptides (Steinman et al., 2003). A significant proportion of the pregnancies and especially those involving a male fetus prior to SRM are associated with obstetric and neonatal complications (Nielsen et al., 2010b). These complications are associated both with an increased transfer of fetal cells into the maternal circulation (Lo et al., 1999; Leunget al., 2001; Zhong et al., 2001; Khosrotehrani et al., 2003) and with increased production of inflammatory cytokines systemically or locally in uterus (Gerber et al., 2005; Girardi et al., 2006; Germain et al., 2007). Circumstances for sensitization of the adaptive immune system against fetal or trophoblast antigens are accordingly often present in the first ongoing pregnancy of SRM patients and may be further increased in patients who are genetically predisposed to anti-HY immune responses, e.g. those carrying the H-Y-restricting HLA class II alleles. A recent study found some healthy female donors sensitized against H-Y while others were H-Y tolerant (van Halteren et al., 2009). Murine studies have demonstrated tolerogenic mechanisms involved in pregnancy. Female mice sensitized (recognizing and destroying) known paternal antigens before pregnancy became tolerant to the same antigens during pregnancy (Zhang et al., 1992; Tafuri et al., 1995; Jiang and Vacchio, 1998). Suppression of these tolerogenic mechanisms may leave the H-Y immunization harmful to male fetuses as demonstrated when blocking TREG in a murine study (Kahn and Baltimore, 2010). TREGs are found to be decreased by numbers in addition to be functionally deficient in RM patients (Arruvito et al., 2007). Down-regulation of the CD4+CD25brightTREG and CTLA-4 mRNA expression in the peripheral and decidual lymphocytes are described in human miscarriage compared with normal pregnancy. CTLA-4 are thought to inhibit TCTL responses resulting in T-cell anergy (Chambers et al., 2001). H-Y-restricting HLA class II alleles present H-Y peptides to maternal CD4+ T cells. These T cells provide help for the efferent phase of the anti-HY immunity, induction of CD8+-mediated cytotoxicity against a series of minor histocompatibility antigens on the trophoblast, secretion of inflammatory cytokines and production of anti-HY and HLA antibodiesall factors being potential harmful for the pregnancy. Figure 1 illustrates the hypothesis of the pathogenesis of SRM after the birth of a boy with a possible scenario at the feto-maternal interface in a pregnant woman with a history of SRM who carries the H-Y-restricting HLA class II allele, HLA-DRB1*15. The figure suggests a series of events ranging from the afferent phase of sensitization where the fetal H-Y protein DDX3Y is processed to the SKGRYIPPHLR peptide by maternal macrophages that can present the peptide/HLA-DRB1*15 complex to maternal CD4+ T cells. These TREG initiate the efferent phase of the anti-H-Y immune reaction; activate TCTL, secretion of inflammatory cytokines, and production of H-Y and HLA-antibodies.
The H-Y hypothesis in SRM: Strengths, limitations and perspectives
Strengths

The majority of our studies testing the H-Y hypothesis in SRM patients is based on patients referred to the Danish Recurrent Miscarriage Clinic. This is a national, tertiary clinic that investigates, treats and conducts research in recurrent pregnancy losses. The clinic is publicly funded and there is access for all eligible patients. There may be a tendency for earlier referral if living closer to the clinic and of higher socioeconomic status (data not available). Such selection biases are not likely to influence the results reported here. Later referral results in more prior miscarriages so all analysis are adjusted for the number of prior losses. The H-Y hypothesis in SRM is based on data from a large group of patients that have undergone a standardized investigation program. All pregnancies before referral are confirmed and described in records from hospitals or practitioners and follow-up after referral is almost 100%. Pregnancy outcome after referral according to the sex of the firstborn was almost identical in two independent cohorts of SRM patients referred before and after 2000 (Nielsen et al., 2008a).

The observed association between a firstborn boy and subsequent low birthweight and obstetric and neonatal complications in SRM patients was also detected, although attenuated, in register studies based on the background populations (Magnus et al., 1985; Nielsen et al., 2008b,2010a; Mortensen et al., 2011) emphasizing the credibility of the epidemiologic results in SRM patients. Our immunological studies testing the H-Y hypothesis have focused on antibodies. Antibody testing has methodological advantage compared with investigation of cellular factors. The laboratory methods are often more reproducible because they are simpler, and the presence of antibodies in the peripheral blood is expected to affect immunological conditions in the uterus. Full-length recombinant H-Y proteins and the H-X homologs were used as targets in the anti H-Yantibody ELISAan assay that has been validated in previous studies of GvHD (Miklos et al., 2005). Only ELISA responses that were specific to H-Y proteins were considered a measure of HY antibodies.
Limitations

One weakness in the studies of the negative prognostic impact of a firstborn boy in SRM patients is the fact that a substantial proportion of the patients were treated with intravenous immunoglobulin (IvIg), which may decrease the miscarriage risk (Christiansen et al., 2002). However, the sex of the firstborn child was not a determinant for offering IvIg treatment to the patients. Furthermore, the two cohorts that were monitored prospectively showed similar higher live birth rate among patients with a firstborn girl compared with a boy in spite of the fact that in the first cohort 15% were offered IvIg, and in the second cohort 45% (Nielsen et al., 2008a). Except for women with complicated pregnancies prior to the miscarriages (Nielsen et al., 2009) and women with severe recurrent placental abruptions (Nielsen et al., 2007) our immunogenetic studies found no differences in H-Y-restricting HLA class II allele frequency in women with SRM and firstborn boys compared with firstborn girls. If H-Y-restricting HLA class II alleles truly confer susceptibility to RM they are expected to show increased frequencies in this population. The heterogeneous background of SRM may explain the observed similar frequencies. Patients suffering RMs due to aneuploid embryos are of course not expected to exhibit any HLA associations. Women with a previous aneuploid miscarriage display a better prognosis than those with previous euploid miscarriage (Boue et al., 1975; Stephenson et al., 2002). Accordingly, patients with immunological may manifest themselves most clearly through a poor prognosis in the next pregnancy. The conclusions from our immunological studies may be limited due to relatively low numbers. In support of the H-Y hypothesis these studies found the presence of HLA and H-Y antibodies to be associated with a lower live birth rate and a low male:female ratio in subsequent live births, respectively (Nielsen et al., 2010c, d), consistent with a direct cytotoxic capacity as in allogeneic hemotopoitic stem-cell transplantation (Miklos et al., 2005). The frequency of H-Y antibody positivity did, however, not differ significantly between SRM patients with a firstborn boy compared with a girl and neither did H-Y antibody positive pregnant patients have a significantly higher miscarriage rate. Both these observations seem to be at odds with the H-Y hypothesis. However, the frequency of H-Y antibodies was 11% higher in SRM patients with a firstborn boy and the subsequent miscarriage rate was 13% higher among H-Y antibody positive patients. Future adequately powered studies need to address these issues.
Perspectives and future studies

The observation that the sex of the firstborn child and presence of particular HLA class II alleles and HLA and H-Y antibodies characterize SRM and subsequent pregnancy outcome provides valuable insight in the pathogenesis of many cases of SRM and may also explain some obstetric and perinatal complications in the background population. However, none of the clinical and immunological risk factors reviewed has sufficiently high specificity for predicting future obstetric and perinatal complications to be introduced in clinical practice. It is therefore important further to identify the putative immunological interactions responsible for the anti-H-Y related pregnancy complications.

HLA antibodies have until now never convincingly been found to play any pathologic role for outcome in normal women or patients with RM but have instead been regarded a result of previous ongoing pregnancies (Sargent et al., 1988; Coulam, 1992). It remains to be investigated whether HLA antibodies in SRM patients are the direct cause of the increased miscarriage risk or whether it is an epiphenomenon that reflects a series of immunological disturbances possibly based on prior abnormal transfers of fetal cells to the maternal immune circulation. It also remains to be determined whether anti H-Y antibodies are directly causing implantation failure and early miscarriage or whether they are epiphenomenons to causal cellular reactions. Most studies of antiH-Y immunity as a cause of GvHD have focused on cellular reactions. However, studies on cellular immunologic reactions in peripheral blood in pregnant women do not necessarily reflect the conditions at the feto-maternal interface and in case of H-Y-specific T-cells, the frequency of these cells is very low (Mutis et al., 1999) requiring large volumes of blood drawn. We did, however, attempt to analyze the presence of functionally different types of H-Y-specific CD8+ T cells, i.e. TREG and TCTL in five SRM patients (after ethical approval and informed consent) referred to the clinic and prior to their next pregnancy attempt. We used exactly the same approach as van Halteren et al. (2009) enabling us to use their healthy donors as controls. Outgrowth of H-Yspecific T cells in all five samples failed whereas van Halteren et al. (2009) had a 50% outgrowth rate. Whether the lack of H-Y-specific CD8+ T cells was due to bad luck due to few samples or whether it truly reflects a lower than expected frequency of H-Y-specific T cells in SRM patients needs to be tested in larger studies. Future studies should aim at exploring how maternal tolerogenic mechanisms against H-Y antigens are established. The nature of CD4+TREG is an obvious candidate. A recent murine study found a portion of the maternal immune response to fetal antigens to be TREG in nature. Depletion of the TREG resulted in a lower fraction of live male offspring and a selective reduction in weight of the surviving males (Kahn and Baltimore, 2010). It also remains to be explored how or whether the initially specific anti-H-Y response suggested to develop after the birth of boys in SRM may also affects subsequent female fetuses. Although male fetuses seem to be at the highest risk of demise after a previous birth of a boy in SRM, female fetuses also seem to be in a higher than normal risk of miscarriage. Determinant spreading, the phenomenon that an initially specific immunological reaction with time spreads to be directed against several related proteins is a recognized and common feature in autoimmune diseases (Lehmann et al., 1993; Ott et al., 2004). A spread of an initially H-Y-specific reaction to closely related H-X analogs carried by females might be responsible for the possible impact on pregnancies with girls. It is also possible that anti-HY responses through production of Th1 cytokines, such as TNF-, could be one of the mechanisms inhibiting TREG cells with specificity for non-sex-specific trophoblast antigens thereby increasing the risk of female fetuses miscarriage too.
Previous SectionNext Section

Conclusion
This review has critically discussed the basis for and the testing of the H-Y hypothesis that aberrant H-Y immunity is a causal factor for SRM. A series of observations lend support to the proposed hypothesis. Sex ratios prior and subsequent to SRM suggest that birth of boys predispose to SRM and male fetuses are more likely to be miscarried. Obstetric complications are more common in pregnancies prior to SRM potentially sensitization the mother against fetal antigens. Maternal carriage of HLA-class II alleles presenting H-Y antigens to the immune cells is associated with a reduced live birth rate and with obstetric complications in surviving pregnancies in SRM patients with a firstborn boy. HLA and anti H-Y antibodies are increased in SRM patients compared with controls. Presence of these antibodies in early pregnancy is associated with a lower live birth rate and a low male:female ratio in subsequent live births, respectively. Births of boys are also associated with subsequent obstetric complications in the background population. These first results need to be confirmed by other RM centers, and collaborative studies across medical specialties and national borders are needed independent testing our hypothesis. This can have implications for our insight into the pathogenesis of RM but may also add to our understanding of maternalfetal interactions during both normal and pathological pregnancies.

Previous SectionNext Section

Funding
The University Hospital Copenhagen Rigshospitalet rewarded a PhD-grant for H.S.N.

The Author 2011. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email: journals.permissions@oup.com Previous Section

References

SNPs in toll-like receptor (TLR) genes as new genetic alterations associated with congenital toxoplasmosis?
W. Wujcicka, J. Wilczyski, and D. Nowakowska
Author information Article notes Copyright and License information

Go to:

Abstract
Nearly 40 % of pregnant women are infected with Toxoplasma gondii. Primary infections in pregnant women result, in approximately 3050 % of patients, in transmission of T. gondii through the placenta to the fetus and then in congenital infections with severe, sometimes fatal course. Studies still do not provide sufficient data on the genetic bases of the immunity in fetuses, newborns, and infants with congenital toxoplasmosis. Previous research showed the contribution of toll-like receptors (TLRs) to non-specific immunity against T. gondii invasion, observed in T. gondii-infected animals, especially mice. So far, the activity of TLRs in defense against T. gondiiinfections was observed particularly for TLR2, TLR4, and TLR9 molecules. Differential TLR activity associates with both cell types, including a variety of placental cells and stage of pregnancy. Several single-nucleotide polymorphisms (SNPs) residing in three genes encoding these receptors were reported as significant genetic modifications of TLRs associated with different pregnancy disorders. Despite those data, genetic alterations of TLRs which have contributed to innate immune response against T. gondii infections are still not precisely described. In this article, we present reasons for the research of the plausible role of SNPs residing in TLR2, TLR4, and TLR9 genes in congenital toxoplasmosis development. Go to:

TLRs contribute to Toxoplasma gondii infections

Toxoplasmosis is one of the most frequent pregnancy infections transmitted from mother to child and the major cause of perinatal morbidity and mortality [13]. In Poland, nearly 40 % of pregnant women are infected with Toxoplasma gondii [4, 5]. Meiosis occurred in oocysts of T. gondii and resulted in the production of many different genotypes of parasitic sporozoites; however, three predominant distinct clonal lines (virulent strain I and non-virulent strains II and III) were found in Europe and North America [1, 6, 7]. Primary infections are dangerous in

pregnant women, occurring usually asymptomatically, resulting, in approximately 3050 % of patients, in transmission of T. gondii through the placenta to the fetus. As a result, these infections cause, in the fetuses and newborns, with immature immune systems, congenital infections with very severe, sometimes fatal course [3, 5, 8]. So far, extensive investigations aimed to describe mechanisms of congenital infections and related immune responses. However, current studies do not provide sufficient data on the associated alterations and genetic background of the innate immunity to T. gondii occurring in newborns and infants congenitally infected with this parasite. Three of ten human toll-like receptors (TLRs), TLR2, TLR4, and TLR9, have been reported to play roles in the recognition of T. gondii, course of infections with this parasite, and also in pregnancy progression and disorders. So far, studies clearly confirmed the essential role of TLR/MyD88 signaling in non-specific antimicrobial immune response to T. gondii infections. Many previous works reported the crucial role of IL-12 and of IFN-, the major regulators of IFN- production, by NK and T cells and cellmediated immunity to T. gondii, respectively ([9, 10], reviewed in [11, 12]). The high expression levels of these two cytokines are stimulated after primary infection with T. gondii [11]. In mice, TLR11 was reported as the main regulator of IL-12 production following T. gondii infection ([11, 13], reviewed in [14]). However, TLR11 in humans is only a non-functional pseudogene [11, 15]. Hence, in this study, we described the role of other receptors from the TLR family in the defense against T. gondii. We suppose that genetic modifications, especially of TLR2, TLR4, and TLR9, might play a plausible role in congenital toxoplasmosis development. To argue the major role of TLRs in innate immune response to T. gondii, we described the most recent studies showing their involvement in immunity to this parasite. Figure 1 illustrates the activities and interactions of TLRs and other molecules reported to be involved in the immune response to T. gondii infection.

Fig. 1

Toll-like receptor (TLR) signaling pathways involved in immune response to Toxoplasma gondii infections. Molecules considered to be crucial during T. gondii infection are underlined and surrounded by a thick continuous line The latest study performed to describe the method of action of lipoxin- and L-kynurenineinduced SOCS2 in the pro-inflammatory response during infection also with T. gondii showed the engagement of both mediators in the inhibition of TLR/MyD88, TLR/TRIF, IL-1R/MyD88, and CD40/CD154 signaling pathways [16] (Fig. 1). Other studies reported that TLR response to T. gondii is promoted by the isoform of the diacylglycerol (DAG) kinases family (DGK), expressed in macrophages (M) and dendritic cells (DC) via a pathway involving the inhibition of PI3K [17] (Fig. 1). DGK deficiency resulted in impaired IL12 and TNF- production after in vitro and in vivo TLR stimulation, elevated resistance to endotoxin shock, and increased susceptibility to T. gondiiinfection [17]. Studies which aimed to describe the role of DC in the immune response to T. gondiishowed that mice with targeted inactivation of MyD88 in these cells, but not in macrophages or neutrophils, were highly susceptible to the T. gondii infection [18]. Subsequent results indicated the crucial role of TLR recognition by DC in the creation of a rapid type 1 innate immunity to T. gondiiand to prevent acute mortality [18]. Murine CD11c+Gr-1+ DC infected with T. gondii, in contrast with non-infected cells, had no ability for ex vivo TLR stimulation [19]. Then, the observed results suggested that plasmacytoid DCs (pDCs) were utilized by T. gondii as Trojan horses in the early infection, with suppressed cytokine effector function, and exploited to disseminate the parasite within the host [19]. Earlier studies also reported the important cellintrinsic role of MyD88 in the activation of neutrophils, macrophages, DC, and T cells during immunity to T. gondii [11, 12]. However, the function of MyD88 in neutrophils, macrophages, and monocytes has to be determined in detail. MyD88/ mice orally infected with T. gondii failed to control the parasite and succumbed within 2 weeks of infection, but the i.p. vaccinations of these mice with avirulent T. gondii uracil auxotroph induced strong IFN responses and protective immunity to the invasion of high-virulence T. gondii strains [20]. Hence, the results of this study confirmed that MyD88 is required to control T. gondii infection, but the adaptive immunity might be induced without the involvement of TLR adaptor molecule [20]. As MyD88 might be activated by almost all TLRs except TLR3, recent investigations also aimed to identify particular receptors crucial for the initial recognition of T. gondii. One of the latest

studies showed that single nucleotide-sensing TLR3, TLR7, and TLR9 have no function in controlling the initial activation of innate immune response and host resistance to T. gondii infection in the 3d mice with a non-functional UNC93B1, a critical mediator of the translocation of the three investigated TLRs from ER to endolysosomes [21] (Fig. 1). However, 3d mice were extremely susceptible to infection with T. gondii. Therefore, the likely combined action of nucleotide-sensing TLR3, TLR7, and TLR9 in host defense against T. gondii was suggested [21]. Studies of 264 healthy Russian Karelian children performed to analyze the influence of the gene environment effect of T. gondii, Helicobacter pylori, CD14 159C>T, and TLR4 +1896A>G polymorphisms on the total serum IgE showed no association between T. gondii or H. pyloriseropositivity, or CD14 and TLR4 polymorphisms with IgE levels [22]. However, the constructed multiway analysis of variance (ANOVA) model showed the influence of CD14 159 allele T and H. pylori antibodies status on the serum total IgE [22]. Go to:

The role of TLR2, TLR4, and TLR9 in Toxoplasma gondii infections

Among TLR characteristics for human cells, the activity in sensing T. gondii was especially determined for TLR2, TLR4, and TLR9 molecules [2326]. In the transcriptome-profiling study performed to identify genes synergistically up-regulated by IFN- and TNF in macrophages, several proinflammatory cytokines and also TLR agonists were reported to stimulate increased expression of newly identified immunoresponsive gene 1 (IRG1) transcript [27]. The expression of IRG1 transcript was up-regulated after the stimulation of TLR2, TLR4, and TLR9 by LTA, lipopolysaccharide (LPS), and CpG, respectively. As IRG1 expression was not regulated by TLR3 stimulated by poly I:C, which is not signaled via MyD88, the molecule involved in the signaling of TLR2, TLR4, and TLR9, the MyD88 molecule, was suggested to be required for the induction of IRG1 [27] (Fig. 1). In other work, TLR2-deficient mice infected with 300 cysts of T. gondii died within a period of 10 days; however, all of them survived when the parasite dose was lower [28]. In T. gondiiinfection cases, TLR2 participated in macrophages activation and regulation of parasite-induced NO production [28]. Another study showed that iNOS/ mice were able to control the acute T. gondiiinfection, in which additional effector mechanisms were possibly

regulated by TLR2 [29]. This receptor was also reported to regulate the T. gondii-induced production of the neutrophil-attracting chemokine CCL2 [30]. Experiments with the vaccination of T. gondii-infected mice with T. gondii-derived heat shock protein 70 (T.g.HSP70) gene showed its involvement in DC activation and Th1 polarization observed at draining lymph nodes (dLN) from wild type and TLR2-deficient mice, but not TLR4-deficient individuals with B6 background [31]. Early Th1 polarization was induced at the dLN of mice by the T.g.HSP70 gene vaccine through the TLR4/MyD88 signal pathway (Fig. 1). At an acute phase of toxoplasmosis, the T.g.HSP70 gene vaccine also limited the copy number of T. gondiiin the mesenteric LN of WT, TLR2-deficient, and TRIF-deficient mice, but for neither TLR4-deficient nor MyD88deficient mice, indicating the involvement of TLR4 in the vaccine effect at an acute phase of infection. What is more, TLR4 had activity in the determination of the T. gondii cyst number of the brain at a chronic phase of toxoplasmosis, observed also at 8 and 12 weeks post-infection [31]. TLR4-mediated macrophage activation was observed during T. gondii andLeishmania donovani infection, although this response was differentially regulated by progesterone via the glucocorticoid and progesterone receptors [23]. Another study showed that purified T. gondii glycosylphosphatidylinositols (GPI) triggered TLR4 signaling pathways as well [24]. Mice lacking both TLR2 and TLR4, but not TLR2 alone, had completely abrogated production of TNF in the response of T. gondii GPI [24]. However, mice deficient in TLR4 were only a little more susceptible to the T. gondii infection [24]. A significant role of TLR4 was reported in T. gondiiinfection observed in the small intestine [32]. TLR9-deficient mice infected orally with T. gondiiwere relatively resistant to the ileitis and had reduced Th1 response to the parasite [25, 26]. TLR9 deficiency in mice infected orally with T. gondii resulted in increased susceptibility to the infection and a 50 % reduction in IFN- production [26]. Then, the results proved that the TLR9 contribution coordinated both innate and adaptive immune response to T. gondii. However, the crucial indirect role in the stimulation of immunity to T. gondii is possibly characteristic for commensal gut bacteria rather than the parasite itself [25, 33]. Go to:

The activity of TLRs during pregnancy

The expression of TLR genes was observed in immune and also in non-immune cells such as trophoblasts, decidual cells, and amniotic epithelium [3436]. The expression profile of TLR was

associated with cell types, stage of pregnancy, and response to microorganisms ([34, 3638], reviewed in [39, 40]). Studies performed for paraffin-embedded sections of endometrium and deciduas from first and second trimester elective terminations and third trimester normal deliveries showed alterations in the TLR4 expression levels between particular tissue samples [34]. Higher immunoreactivity of TLR4 observed in decidual cells compared to interstitial trophoblasts suggested maternally derived cells as the main protectors against Gram-negative bacteria and other harmful signals from severe inflammation associated with or without infection. Particularly, significantly higher TLR4 levels were observed in decidual cells of the first and third trimester compared to second trimester basalis, suggesting that an enhanced inflammatory environment occurred at the maternalfetal interface in the initial stages and end of pregnancy compared to mid-pregnancy [34]. In turn, another study showed the presence of TLR2 and TLR4 proteins expressions from the one-cell stage through the blastocyst stage during murine embryo development [41]. In trophoblast stem cells, the expression of TLR2 and TLR6 but not TLR1 or TLR4 was observed [41]. At the plasma membrane, the expression of only TLR2 molecules was determined. The limited expression of TLRs on trophoblast cells suggested that early trophoblasts might be less able to protect embryos against pathogenic agents compared to differentiated trophoblast cells [41]. The obtained results are consistent with outcomes of another study, which indicated that TLR4 expression and related IL-8 secretion that occurred after LPS stimulation might be assigned, rather, to leukocytes, as it was no longer observed in trophoblast fraction purified from leukocytes [42]. The expression of different antimicrobial peptides and proteins was suggested to be characteristic for various placental cells and resulted from the cooperation between leukocytes and cells of embryonic origin [42]. Functional screening tests of TLRs in human amniotic epithelial cells showed that the expression of TLR5, TLR6/2, and TLR4 occurred in these cells [35]. The activation of TLR5 and TLR6/2 stimulated both IL-6 and IL-8 production and the NF-B signaling pathway, whereas TLR4 activity was associated with reduced viability of analyzed cells and their apoptosis. Hence, differential immune response mediated via specific TLRs might even lead to preterm birth [35]. Other data suggested a contribution of TLR ligands to the direct activation of decidual NK cells and also NK cell lines observed in the presence of proinflammatory cytokines [43]. In human decidual mononuclear cells (DMNC), LPS, known as the TLR4 ligand, induced IFN- and TNF- production, from which the former was IL-2- and IL-12-dependent, while the latter was independent [43]. Hence, further studies

were suggested to be necessary to determine the mechanisms driving the synergistic effect of TLR and cytokine signaling [43]. Research of cultured cells of term placenta, including cytotrophoblast- and syncytiotrophoblast-rich cells, showed the expression of TLR2, TLR3, TLR4, TLR5, TLR6, and TLR9 genes [36]. In first-trimester primary trophoblasts and trophoblast cell lines, the expression of TLR1, TLR2, TLR3, and TLR4, but not TLR6, was observed [37, 44]. However, the last gene was expressed in third-trimester trophoblasts, which suggested its regulation in a temporal manner [36, 37]. The TLR4 expression level was higher in the term compared to the first-trimester placentas [45]. Placentas from early pregnancy were reported as being less responsive to pathogen agents than term tissues. TLR2 and TLR4 transcripts were identified in villous cytotrophoblast and extravillous trophoblast, but not in syncytiotrophoblasts in the first-trimester placenta [37]. It also suggested the regulation of TLR in a spatial manner. Then, it seemed that pathogens might be dangerous to the fetus only when the TLR-negative outer trophoblast layer is interrupted and infecting agents enter either the placental villous or the decidual compartments [37, 44]. The level of TLR1TLR10 transcripts was also investigated in term human placentas collected in the absence of labor (elective cesarean sections; ECS) and after the completion of labor (normal vaginal delivery; NVD) [38]. TLR1 TLR10 transcripts were observed in all placentas; however, NVD tissues produced higher levels of TNFA in response to TLR4 (LPS) and to TLR7/8 agonist (resiquimod) than ECS placenta explants. Significantly elevated levels of TLR2 and TLR5 correlated with labor. Hence, this data suggested the role of TLR in parturition [38]. Studies of mRNA expressions of TLR2, TLR3, TLR4, and TLR9 in the uterus, cervix, and placenta of non-pregnant and across gestation CD-1 mice showed significantly elevated levels in pregnant uterine and cervical tissues [46]. The differential levels of TLR mRNA expressions were observed between the uterus, cervix, and placenta with significantly down-regulated TLR4 gene expression in the placenta [46]. The data suggested that the innate immune system was an active and dynamic system during gestation, with the altered expression of TLRs at the maternalfetal interface playing a crucial role in the pathogenesis of adverse pregnancy outcomes. Despite data confirming the involvement of TLR2, TLR4, and TLR9 receptors in the non-specific immunity to T. gondii, their contribution to toxoplasmosis development, and function within pregnancy, detailed mechanisms driving the activity of these receptors in T. gondii congenital

infections are not known. So far, previous research showed the participation of several singlenucleotide polymorphisms (SNPs) of TLR2, TLR4, and TLR9 in various diseases and pregnancy disorders [4749]. Go to:

SNPs in the TLR2 gene associated with pregnancy disorders

In a group of 200 cord blood samples obtained from 72 atopic and 128 non-atopic mothers, 12 SNPs located in TLR1, TLR2, TLR4, TLR6, and TLR10 were genotyped to describe their possible influence on T-regulatory cells required for keeping immune responses in balance and roles played in atopic diseases [50]. In neonates, the presence of AA genotype of the TLR2 promoter rs4696480 SNP associated with the increased expression of FOXP3 and Treg marker genes GITR and LAG3, and also the secretion of TH2 cytokines and TNF- in case of maternal atopy, with Tregs diminished without maternal atopy. The occurrence of GG genotype of TLR2 rs1898830 correlated with Treg marker genes decreasing with and increasing without maternal atopy [50]. The screening of common TLR2(2258 G>A) and cosegregating TLR4 (1063A>G and 1363C>T) SNPs in 94 women with preeclampsia and 176 healthy pregnancy controls showed the associations of these SNPs with early-onset but not lateonset preeclampsia [47]. Then, three TLR SNPs lowering thresholds for early-onset and severe pregnancy disease, but not late-onset or mild pregnancy disease, was described [47]. In turn, a study of 288 vaginal samples obtained from 144 women during both the first and second trimesters of pregnancy performed to determine the correlation of 34 SNPs residing in nine genes involved in TLR-mediated and -related sensing and regulation of pathogens with the vaginal carriage of Gram-positive anaerobes Gardnerella vaginalis and Atopobium vaginae showed no association between gene polymorphisms and bacterial vaginosis (BV) [51]. No significant correlation was determined for TLR2 15607A>G, 1350T>C, 2258G>A, and 2029C>T SNPs [51]. The lack of a relationship betweenTLR2 polymorphisms and BV occurrence is consistent with the results of another study, performed to determine the influence of four TLR2 SNPs [rs18988 (15607 A>G), rs4696483, rs7656411, and rs1337)] on the cervical levels of pro- and anti-inflammatory cytokines and their association with BV [52]. In another study, analyses of the relevance of polymorphisms located in genes associated with innate immunity performed for a large cohort of preterm very-low-birth and term-born infants and their

mothers showed no significant correlation between the presence of TLR2 2258G>A SNP and the occurrence of intrauterine infections as the cause of preterm birth [53]. Another study also showed that comparable frequencies of the mutated 2258G allele of the TLR2 gene occurred in both pregnant women with preterm labor and healthy individuals [54]. In turn, another investigation reported that infants with two polymorphic TLR2 alleles (16934TA/AA and 2258GA/AA) had, importantly, shorter gestational ages [55]. Go to:

SNPs in the TLR4 gene associated with pregnancy disorders

For the TLR4 gene, 1063A>G and 1363C>T SNPs were the most commonly studied. Real-time polymerase chain reaction (PCR) assays followed by high resolution melt (HRM) analysis, performed to simultaneously identify SNPs from TLR4 (1063A>G) as well as other genes IL6, IL1, and IL12RB involved in the immune response showed no association of these polymorphisms with preterm birth of infants in Montevideo, Uruguay [48]. However, in the same population, fetuses carrying the 1063A>G SNP were both severely premature and had premature rupture of membranes (PROM) at the same time [56]. In a Canadian cohort which included mainly preterm infants, this TLR4 SNP occurred at a significantly lower frequency in infants without bronchopulmonary dysplasia (BPD) compared to those with diagnosed BPD. However, no significant association was observed between TLR4 genotypes and prematurity [57]. In turn, the research of a Finnish population showed the presence of the 1063A>G SNP in both infants and mothers and its correlation with preterm labor [58]. The frequency of TLR4 1063A/G alleles and their association with preterm labor might be population-specific. It seems that further determination of the contribution of the 1063A>G SNP to preterm labor is required. In contrast toTLR2 polymorphisms, two SNPs located in the TLR4 sequence (rs1554973 or rs7856729) interacted significantly with IL-1R2 rs485127 and correlated with the concentration of IL-1 in 188 African American and European American women [59]. The study showed both TLR4 and IL1-R2 genotypes as possible significant markers of cervical cytokines IL-1 or IL-1 concentrations and involvement in disparity in pregnancy outcomes. However, the observed associations were no longer significant after correction for multiple testing in European or African Americans [59]. An earlier study reported a genetic association between TLR4 rs1554973 and the level of cervical cytokine IL-1 characteristic for European but

not African Americans, which was more significant in women with BV compared to those without BV [52]. These results are consistent with the observation that selection on TLRs varies between human populations [60]. In another study, the carriage of TLR4rs1554973 SNP strongly associated with chorionic plate inflammation both in mothers and their singleton fetuses [61]. Four SNPs of TLR4 [1063A>G (rs4986790), 1363C>T (rs4986791), rs6478317, rs10759932] were reported to affect T-cell regulation [50]. Two common 1063A>G (rs4986790) and 1363C>T TLR4 SNPs correlated with increased IL-13 secretion in core blood mononuclear cells (CBMC) obtained from umbilical veins after delivery and stimulated with Dermatophagoides pteronyssinus. CBMC with TLR4 rs6478317 (GG) had higher IL-12 secretion observed after stimulation with lipid A (LpA). 1063A>G (rs4986790) and 1363C>T TLR4 SNPs were associated with a higher risk for atopic dermatitis [50]. Similarly to TLR2 SNPs, no TLR4 polymorphisms (rs1927914, 1063A>G, 1363C>T) were associated with BV in 288 vaginal samples obtained from 144 women during both the first and second trimesters of pregnancy [51]. However, another study showed that the 1063A>G SNP, disturbing response to LPS, was correlated with an increase in the vaginal pH and levels of Gardnerella vaginalis and anaerobic Gram-negative bacteria [62]. In a study of the occurrence of the 1363C>T SNP in pregnant women with preterm labor and a control group, a lower frequency of heterozygous CT genotype and mutated T allele was observed in the study group [54]. Hence, this SNP might play a protective role against PROM in pregnant women. However, the proposed hypothesis has to be further investigated in a larger group of patients with PROM [54]. In Plasmodium falciparum-infected primiparous Ghanaian women, the 1063A>G SNP elevated the risk of low birth weight in term infants 6-fold and the risk of maternal anemia 5-fold [63]. The obtained results suggested a correlation of the 1063A>G SNP with a clinical picture of malaria during pregnancy. However, this mutation did not associate with preterm labor [63], whereas the 1363C>T SNP was, importantly, less common in women with BV compared with individuals without BV [64]. Go to:

SNPs in the TLR9 gene associated with diseases and ocular toxoplasmosis
The accumulated data argued also for the association of different TLR9 SNPs with a variety of disorders, along with toxoplasmic retinochoroiditis, cervical cancer, mother-to-child transmission (MTCT) of human immunodeficiency virus type 1 (HIV-1), increased risk of low

birth weight in infants, risk of maternal anemia, and a clinical picture of malaria in pregnancy [49, 63, 65, 66]. The 1635A>G SNP residing in the TLR9 gene was reported to correlate with toxoplasmic retinochoroiditis in Brazil [49]. In the studied population, ocular toxoplasmosis was associated with allele C at 1635A>G [odds ratio (OR) 7; 95 % confidence interval (CI) 1.630.8), which was at a frequency of 0.424 %, similar to that observed in European populations. The determined correlation suggested that direct interaction between T. gondii and TLR9 might trigger proinflammatory responses and, hence, led to severe pathologies such as ocular disease, associated with this infection in Brazil [49]. A study performed for the Polish population showed the association of TLR9 1635A>G and also rs187084 polymorphisms with cervical cancer and the possible significance of these SNPs as risk factors of cervical cancer development [65]. In turn, another work showed the role of the c.4-44G>A and c.1635A>G SNPs in MTCT of HIV-1 [66]. While neither of the two SNPs were associated with a risk of HIV-1 infection, the [G;G] haplotype correlated with a higher risk of MTCT of HIV-1 after adjustment for maternal viral load [66]. A study performed for P. falciparum-infected primiparous Ghanaian women reported a correlation of the 1486T>C SNP residing in the TLR9gene, just as in the case of the 1063A>G SNP in the TLR4 gene, with an elevated risk of low birth weight in term infants and the risk of maternal anemia and, therefore, with a clinical picture of malaria in pregnancy [63]. Go to:

Concluding remarks
It is important to investigate molecular mechanisms involved in Toxoplasma gondii congenital infections more precisely. It seems necessary to identify, in clinical specimens such as blood, amniotic fluid, cerebrospinal fluid, and the placenta, the genetic alterations of TLR2, TLR4, andTLR9 genes accompanying the development of these infections and to describe the clinical significance of single-nucleotide polymorphisms (SNPs) residing in these genes in regard to congenital toxoplasmosis. TLR genes were suggested to play roles in toxoplasmosis development, progression of pregnancy, and various disorders of gestation. One possible modification influencing gene function, which was also confirmed in case of TLR2, TLR4, and TLR9 genes in accordance to different diseases and pregnancy disorders, is the occurrence of SNPs within the genetic sequence. Therefore, the detailed description of the occurrence and the role of SNPs located in these three genes of the TLR family might pose a significant direction of research to

determine new genetic alterations involved in immunity against T. gondii infections in pregnant women, their fetuses, and newborns.

NALP1 Influences Susceptibility to Human Congenital Toxoplasmosis, Proinflammatory Cytokine Response, and Fate of Toxoplasma gondiiInfected Monocytic Cells
William H. Witola,1, Ernest Mui,1 Aubrey Hargrave,1 Susan Liu,1 Magali Hypolite,2 Alexandre Montpetit,3 Pierre Cavailles,2 Cordelia Bisanz,2 Marie-France Cesbron-Delauw,2, Gilbert J. Fourni,4,5, and Rima McLeod1,*
Author information Article notes Copyright and License information

This article has been cited by other articles in PMC.

Go to:

ABSTRACT
NALP1 is a member of the NOD-like receptor (NLR) family of proteins that form inflammasomes. Upon cellular infection or stress, inflammasomes are activated, triggering maturation of proinflammatory cytokines and downstream cellular signaling mediated through the MyD88 adaptor. Toxoplasma gondii is an obligate intracellular parasite that stimulates production of high levels of proinflammatory cytokines that are important in innate immunity. In this study, susceptibility alleles for human congenital toxoplasmosis were identified in the NALP1 gene. To investigate the role of the NALP1 inflammasome during infection with T. gondii, we genetically engineered a human monocytic cell line for NALP1 gene knockdown by RNA interference. NALP1 silencing attenuated progression of T. gondii infection, with accelerated host cell death and eventual cell disintegration. In line with this observation, upregulation of the proinflammatory cytokines interleukin-1 (IL-1), IL-18, and IL-12 upon T. gondii infection was not observed in monocytic cells with NALP1 knockdown. These findings suggest that the NALP1 inflammasome is critical for mediating innate immune responses to T. gondii infection and pathogenesis. Although there have been recent advances in understanding the potent activity of inflammasomes in directing innate immune responses to disease, this is the first report, to our knowledge, on the crucial role of the NALP1 inflammasome in the pathogenesis of T. gondii infections in humans. The innate immune system plays an important role in sensing pathogens and triggering biological mechanisms to control infection and eliminate pathogens, as well as marshalling the Tand B-cell responses of adaptive immunity. These functions are achieved through engaging an array of germ line-encoded pattern recognition receptors (PRRs) to detect invariant pathogen motifs (26). PRRs are expressed by cells such as monocytes, macrophages, dendritic cells, neutrophils, and epithelial cells, as well as other cells of the adaptive immune system. PRRs

include Toll-like receptors (32), C-type lectins (17), RIG-like helicases (40), cytosolic DNA sensors (11), and members of the NOD-like receptor (NLR) family (21). The NLRs are a large group of cytosolic sensors whose apparent main function is to modulate the expression of the proinflammatory cytokines interleukin-1 (IL-1) and IL-18. There are >23 NLRs encoded in the human genome, some of which are important in regulation and activation of proinflammatory caspases (34). The NLRs have a tripartite structure composed of a C-terminal leucine-rich repeat domain, a central nucleotide-binding oligomerization domain (NOD or NACHT), and a variable N-terminal protein-protein interaction domain that can be either a caspase recruitment and activation domain, a pyrin domain, or a baculovirus inhibitor of apoptosis repeat domain (13). The Cterminal leucine-rich repeat domain functions in ligand sensing and autoregulation, while the Nterminal caspase recruitment or pyrin domain mediates homotypic protein-protein interactions for downstream signaling. The NACHT domain enables activation of the signaling complex via ATP-dependent oligomerization (36). Members of the NLR family assemble into large multiprotein complexes (700 kDa), termed inflammasomes, that can be activated by cellular pathogens (bacteria, fungi, viruses, and protozoa) or stress to engage innate immune defenses (23, 35). Inflammasomes activate caspase-1, a proteolytic enzyme that cleaves and activates the secreted cytokines IL-1 and IL-18 (13, 27). Inflammasomes are also closely linked to pyroptosis, a caspase-1-dependent highly inflammatory cell death process often observed during infection with cytosolic pathogens (36). NALP1 is a member of the NLR family and is the only NLR protein that contains an additional Cterminal caspase activation and recruitment domain (CARD) (23). Upon stimulation, NALP1 is thought to assemble an inflammasome complex consisting of NALP1, apoptosis-associated specklike protein containing a CARD, caspase-1, and caspase-5 (23, 37). The physiological stimulus of this activation has yet to be elucidated, although it is thought that the leucine-rich repeats of NALP1 recognize a danger signal, triggering inflammasome assembly, the activation of caspases 1 and -5, and subsequent processing of IL-1 (19). Toxoplasma gondii is an intracellular apicomplexan parasite that has the ability to infect virtually all warm-blooded animals, with about one-third of the human world population being

seropositive for T. gondii (28). It can cause serious illness in immunocompromised individuals and in those infected congenitally. The hallmark of T. gondii is its ability to induce long-term chronic infections necessitated by specific parasite-host interactions, with conversion from the prolific tachyzoite stage to the quiescent bradyzoite stage through modulation of host cell responses associated with pathway signaling, cytokine production, and antimicrobial effector mechanisms, among others (1). Recognizing that the NALP1 gene is within a robust toxoplasmosis susceptibility/resistance region in the rat genome (Toxo1) (6, 7) that is orthologous to a region in the human genome, we hypothesized that the NALP1 gene might have susceptibility alleles for human congenital toxoplasmosis and could therefore have a significant role in the pathogenesis of Toxoplasma gondiiinfection in humans. In this study, transmission disequilibrium testing (TDT) (18) studies were performed to determine whether there is an association of alleles of the NALP1 gene with human congenital toxoplasmosis. Furthermore, in the present study, we endeavored to elucidate the role of NALP1 in the pathogenesis of T. gondii infection in a human monocytic cell line. We provide evidence that silencing of expression of the NALP1 gene by RNA interference (RNAi) alters the T. gondii-induced expression profiles of certain proinflammatory cytokines and increases the progression rate of T. gondii infection in a human monocytic cell line. Go to:

MATERIALS AND METHODS

Patient cohort and genotyping of NALP1.

Case-parent trios were from the National Collaborative Chicago-Based Congenital Toxoplasmosis Study (NCCCTS), as previously described (18). Briefly, the NCCCTS consisted of 179 North American children (clinically diagnosed with congenital toxoplasmosis) and their parents. Of the 179 infected children, 124 (83%) had clinically confirmed brain calcifications, with or without hydrocephalus and/or retinal lesions, at birth or the time of diagnosis. DNAs were extracted successfully from 149 case-parent samples and genotyped at 23 single-nucleotide polymorphism tags (tag-SNPs) distributed throughout the NALP1 gene. The tag-SNPs were selected from the HapMap project, release 21 (http://www.hapmap.org), using 10 kb of flanking sequence on each side of the NALP1 gene, a minor allele frequency (MAF) cutoff of 5% in CEU, and an r2 threshold of 0.8. The Tagger tool within Haploview (http://www.broad.mit.edu/mpg/haploview/) was used

to select tag-SNPs. Allelic association analysis was performed for the 124 infected children in the cohort with confirmed clinical findings for the eye and/or brain, using UNPHASED (http://www.mrc-bsu.cam.ac.uk/personal/frank/software/unphased/).
Plasmid constructs.

Short hairpin RNA (shRNA) sequences for the human NALP1 gene coding sequence (GenBank accession no. NM_001033053) and the Salmonella enterica tetracycline repressor (TetR) gene coding sequence (GenBank accession no. NC_006856) were designed for cloning into a Gatewayadapted entry vector (Invitrogen). For the NALP1 coding sequence, the sense and antisense shRNA sequences were 5caccAACAGACATGGAGCTCTTAGTGTGCACTTcgaaAAGTGCACACTAAGAGCTCCATGTCT G-3 and 5aaaaCAGACATGGAGCTCTTAGTGTGCACTTttcgAAGTGCACACTAAGAGCTCCATGTCTGT T-3 (the first four nucleotides in lowercase facilitate directional cloning, the sequence in bold uppercase is the target sequence, the subsequent italicized lowercase sequence is for loop formation, and the underlined sequence is the antisense sequence for the target sequence), respectively. For the TetR coding sequence, the sense and antisense shRNA sequences were 5caccAACGGCCGACGCGCAGCTTCGCTTCCTCTGcgaaCAGAGGAAGCGAAGCTGCGCGTCG GCCGTA-3 and 5aaaaTACGGCCGACGCGCAGCTTCGCTTCCTCTGttcgCAGAGGAAGCGAAGCTGCGCGTCGG CCGTT-3, respectively. For NALP1 and TetR, the target sequences were 2,238 to 2,265 bp and 3,360 to 3,389 bp of the coding sequences, respectively. To clone double-stranded shRNA sequences into the entry vector (pENTR/H1/TO), the sense and antisense shRNA sequences were annealed in vitro and ligated directionally into the pENTR/H1/TO entry vector, essentially as described in the manufacturer's user manual (17a). Upon propagation of the recombinant plasmids, the shRNA inserts were confirmed by sequencing. The recombinant plasmids were used in LR recombination reactions to transfer the shRNA into the expression vector pLenti4/Block-iT-DEST following the manufacturer's instructions (17b). The recombinant destination plasmids were propagated inEscherichia coli, extracted, and sequenced to ascertain that the shRNA inserts were correct.
Generation of NALP1- and TetR-expressing lentiviruses.

To produce lentiviral stocks, the pLenti4/Block-iT-DEST vector carrying the shRNA of interest (NALP1 or TetR) and optimized ViraPowerPackaging mix (containing plasmids pLP1, pLP2, and pLP/VSVG, which facilitate viral packaging) were transfected into 293 FT cells following the manufacturer's instructions (Invitrogen). The transfected cells were cultured for 72 h in complete Dulbecco's modified Eagle's medium (DMEM; 10% fetal calf serum, 2 mM L-glutamine, 0.1 mM MEM nonessential amino acids, 1 mM sodium pyruvate, 1% penicillin-streptomycin). The viruscontaining supernatants were harvested by centrifugation and stored in aliquots at 70C. The viral yield was determined by titration following the instructions outlined in the user manual (17c).
Establishment of human monocytic cell lines for stable expression of NALP1 or TetR shRNA.

To establish a human monocytic cell line stably expressing either NALP1 or TetR shRNA, in vitroculture-adapted human monocytes, i.e., MonoMac6 cells (41), were used. MonoMac6 cells were seeded at 4 105/well in a 24-well plate in RPMI medium supplemented with 10% fetal calf serum, 2.05 mM L-glutamine, 1 nonessential amino acids (Sigma), OPI medium supplement (Hybri Max; Sigma), and 1% penicillin-streptomycin. About 250 l of the NALP1 or TetR shRNAexpressing lentiviral stock (3 106 transducing units [TU]/ml) was added to the cultures and mixed. Untransduced cell cultures were maintained as controls. The cells were maintained at 37C with 5% CO2. The following day, Zeocin (Invitrogen) was introduced into the cultures at 100 g/ml to select for transduced cells. The cultures were maintained by changing the medium (with Zeocin) every 2 days continuously for about 3 weeks, when cells resistant to Zeocin started propagating.
Analysis of RNAi-based knockdown of NALP1 gene transcription.

To determine the effect of lentivirus-expressed NALP1 shRNA on the silencing of NALP1 gene expression via RNAi in MonoMac6 cells, triplicate sets of MonoMac6 cells engineered to express either NALP1 or TetR gene shRNA and of wild-type cells were seeded at 1 105 cells per well in a 12-well plate and cultured for 72 h. The cells were harvested by centrifugation, and total RNA was extracted by use of Trizol reagent. Exactly 2 g of RNA for each sample was treated with DNase I to remove residual genomic DNA, and reverse transcription (RT) was performed using SuperScript III first-strand synthesis supermix for quantitative RT-PCR (qRT-PCR) (Invitrogen). The primer set used for the NALP1 gene was NALP1-F and NALP1-R (Table (Table1).1). This set

of primers amplified a 280-bp fragment of the NALP1 open reading frame. Primers for the human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene used for real-time PCR were GAPD-F and GAPD-R (Table (Table1).1). This set amplified a 300-bp fragment of the open reading frame of the human GAPDH gene. Both the NALP1 and GAPDH fragments were synthesized by conventional PCR from cDNA and then gel extracted, and serial 10-fold dilutions were made as templates to generate a quantification standard curve. The real-time PCR mix consisted of 1 l of cDNA template, 1 l of primer mix (500 nM [each]), and 10 l of SsoFast EvaGreen supermix (Bio-Rad), with the final volume made up to 20 l with RNase- and DNasefree water. The cycling conditions included an initial denaturation for 30 s at 95C, 40 cycles of 95C for 5 s, 57C for 5 s, and 60C for 10 s, and a final cooling step to 40C. Cycling was performed using a CFX 96 real-time system (Bio-Rad), and transcript quantities were derived by the system software, using the generated standard curves.

TABLE 1. Primers used in this study

Western blot analysis of NALP1 protein expression.

MonoMac6 cells engineered to express either NALP1 or TetR gene shRNA and wild-type cells were seeded at 4 105 cells per well in a 6-well plate and cultured for 72 h. The cells were harvested by centrifugation, and the pellet was washed twice in phosphate-buffered saline (PBS) and lysed in SDS sample buffer. Equal protein amounts were loaded into wells of a 12% SDSpolyacrylamide gel and fractionated by electrophoresis. The proteins were transferred to nitrocellulose membranes, and immunoblotting was done using either rabbit anti-NALP1 (Sigma) or mouse anti-GAPDH (LifeSpan Biosciences) antibodies as primary antibodies, with conjugated goat anti-rabbit and anti-mouse antibodies as the secondary antibodies. Signal generation was performed using an ECL chemiluminescence kit (PerkinElmer Life Sciences).

Cell viability assay.

To study the effect of NALP1 gene knockdown on the viability of human monocytic cell cultures during infection with T. gondii, wild-type MonoMac6 cells and those genetically modified for the stable knockdown of the NALP1 or TetR gene were seeded in triplicate into 96-well plates at a density of 104 cells/well and cultured in 200 l of supplemented RPMI medium in the presence or absence of T. gondii strain RH parasites at a multiplicity of infection (MOI) of 1:4 (parasite:monocyte) for different times (0, 1, 2, 3, and 4 days). Wells containing parasites only, without MonoMac6 cells, were also set up as controls. A colorimetric assay using the cell proliferation reagent WST-1 (Roche) for the quantification of cell viability was performed on the cultures at different time points by adding 10 l of the WST-1 reagent to each well. After mixing of the samples, the plates were wrapped in aluminum foil and incubated for 1 h at 37C with 5% CO2. Three independent experiments were performed, with triplicate samples for each experiment, and quantification of the formazan dye produced by metabolically active cells was read as the absorbance at a wavelength of 420 nm, using a scanning multiwell spectrophotometer (Spectra Max 250; Molecular Devices).
Preparation and microscopic analysis of Giemsa-stained MonoMac6 smears.

To determine the effect of NALP1 gene knockdown on monocytic cell viability when cells were cultured with or without parasites, wild-type MonoMac6 cells or those genetically modified for the stable knockdown of the NALP1 or TetR gene were seeded into 24-well plates at a density of 105 cells/well and cultured with or without parasites at an MOI of 1:4 (parasites:monocytes) for different times (0, 1, 2, 3, and 4 days). At 0, 1, 2, 3, and 4 days postinfection, 50 l of the thoroughly mixed culture was used to prepare cytospin smears on glass slides. The slides were air dried and fixed in amino acridine, followed by Giemsa staining and light microscopic examination. The number of infected cells per microscopic field was counted using a 20 objective and derived as a percentage of the total number of cells per field. For each sample, a total of 50 microscopic fields per time point were counted, and the data are presented as the mean percentages of infected cells for 50 fields. Furthermore, the average number of parasites in one parasitophorous vacuole of an infected cell was determined for 20 infected cells per field, with a total of 10 fields being counted per time point.
Quantitation of expression of cytokines, BCL2A1, and caspase-1.

To determine the effect of NALP1 gene knockdown in MonoMac6 cells on the expression of the cytokines IL-18, IL-1, IL-12A, tumor necrosis factor alpha (TNF-), and gamma interferon (IFN), as well as that of the BCL2-related protein A1 (BCL2A1) and caspase-1, triplicate sets of wildtype MonoMac6 cells and MonoMac6 cells engineered to express either NALP1 or TetR shRNA were cultured with or without parasites at an MOI of 1:4 for 36 h. The cells were harvested, and either total RNA was extracted for transcript analysis or total protein lysates were prepared for Western blotting of cytokine protein expression. Total RNA was extracted by use of Trizol reagent, followed by first-strand cDNA synthesis as described above. Quantitative real-time PCR was performed on each sample of cDNA to determine the transcript levels. Primers used were designed to amplify a fragment of approximately 300 bp for the coding sequences of IL-18, IL-1, IL-12A, TNF-, IFN-, BCL2A1, and caspase-1 (GenBank accession numbers NM_001562, NM_000576, NM_000882, NM_000594, NM_000619, BT007103, and NM_033294, respectively). Primer sets for IL-18, IL-1, IL-12A, TNF-, IFN-, BCL2A1, and caspase-1 were IL-18-F and IL-18-R, IL-1-F and IL-1-R, IL-12A-F and IL-12A-R, TNF--F and TNF--R, IFN--F and IFN--R, BCL2A-F and BCL2AR, and Casp-F and Casp-R, respectively (Table (Table1).1). The respective gene fragments were amplified by conventional PCR from the cDNA and sequenced to confirm their identity. The real-time PCR mixture and cycling conditions were essentially as described above. The transcript levels of human GAPDH (as described above) and human actin (GenBank accession number NM_001100) were determined for each sample and used for normalization. The primer set for human actin was hACT-F and hACT-R (Table (Table1).1). Western blotting was performed on equal amounts of lysate protein, using antibodies generated against human IL-1 and GAPDH.
Statistical analysis.

Statistical analyses for all in vitro assays were performed using two-tailed Student's t test. For the genetic study, allelic association analysis was performed using conventional TDT to determine the linkage disequilibrium (LD). LD is the nonrandom association of alleles at two or more loci. It is essentially an approximation of the existence of historical recombination between 2 loci. P values were calculated using Haploview (http://www.broadinstitute.org/haploview). P values of 0.05 or less were considered significant for association with disease.

Go to:

RESULTS
Congenital toxoplasmosis associates with NALP1 in humans.

To test the association of NALP1 with congenital toxoplasmosis, we genotyped a total of 23 tagSNPs that were selected in the human NALP1 gene in a North American cohort of case-parent trios comprising 124 congenitally infected children, many with ocular and/or brain disease, using an r2 threshold of 0.8. SNPs with a call rate of >90% and in HWE in the parents were kept for further analysis. Among the SNPs tested, NALP1 rs8081261 and rs11652907 associated with congenital toxoplasmosis (P < 0.00268 and P < 0.02, respectively) (Fig. (Fig.1).1). An etiological variant of SNPs in strong LD with these markers could account for the observed association with susceptibility to this congenital disease.

FIG. 1. Analysis of NALP1 SNPs. The upper diagram shows the positions of genotyped SNPs relative to the intron-exon structure of the gene. The lower diagram shows the LD plots generated in Haploview, using NALP1 gene data from our North American patient cohort. ...
Establishment of human monocytic cell lines stably expressing NALP1 and TetR gene shRNAs.

To generate human monocytic cell lines that stably expressed either NALP1 or TetR gene shRNA for targeted knockdown of the respective gene by RNAi, lentiviral stocks expressing these shRNAs were engineered and titrated and found to be present at approximately 3 106 TU/ml, sufficient to transduce about 1 107 cells at an MOI of 1. The NALP1 shRNA-expressing lentivirus was used for targeted silencing of the endogenous NALP1 gene in monocytic cells, while the TetR shRNA was generated as an off-target control. The design of the promoter region that drives the expression of the target gene shRNA (Fig. (Fig.2A)2A) facilitates tetracycline-based activation of shRNA expression in a cell line expressing a tetracycline repressor protein. However, in the absence of the tetracycline repressor protein, the transduced cell line would constitutively express the target gene shRNA. We therefore first conducted a preliminary study to determine

the effect of constitutive expression of NALP1 or TetR shRNA in MonoMac6 cells lacking the tetracycline repressor expression vector. We found that this (lack of tetracycline repressor protein regulation of promoter activity) did not affect the morphological appearance or viability of the monocytic cells (data not shown). Based on these observations, we then generated stable monocytic cell lines (lacking the tetracycline repressor expression vector) constitutively expressing either NALP1 or TetR shRNA. These cells did not require the use of tetracycline for expression of the shRNA. Stable cell lines were achieved by taking advantage of the presence of an Sh ble gene in the lentivirus shRNA expression plasmid whose product confers resistance to the antibiotic Zeocin. Upon transduction with the lentivirus, monocytic cells were cultured in the presence of Zeocin at 100 g/ml. This resulted in the majority of the cells dying in about 2 weeks, after which Zeocin-resistant cells started propagating and attained a wild-type growth rate after a further 2 weeks.

FIG. 2. Analysis of the effect of NALP1 and TetR shRNAs on expression of NALP1. (A) Illustration of the promoter features of the pLenti4/Block-iT-DEST shRNA expression vector. In the absence of a Tet repressor protein, there is constitutive expression of the ...
Expression of NALP1 shRNA effectively silences endogenous NALP1 gene expression in human monocytic cells.

To assess the effectiveness of the lentivirus-expressed NALP1 shRNA in silencing the expression of NALP1 gene expression in monocytic cells, triplicate sets of wild-type MonoMac6 cells and those engineered to express either NALP1 gene shRNA or TetR shRNA were cultured for 72 h, after which total RNA was extracted and reverse transcribed. Cell viability was the same in the two transduced cell lines and was compatible with the cell viability of the nontransduced cells. Quantitative real-time PCR analysis of the NALP1 gene transcripts normalized to human GAPDH transcripts showed a significant (P < 0.001) reduction (6-fold) in the NALP1 gene transcripts in MonoMac6 cells expressing NALP1 shRNA compared to the wild-type cells as well as those expressing TetR shRNA (Fig. (Fig.2B).2B). To correlate this effect to NALP1 protein levels, protein extracts from cells treated as described above were analyzed by Western blotting, using

anti-human NALP1 and anti-human GAPDH. While there was no notable difference in the detectable protein signals for GAPDH among all samples, there was a significant reduction in the detectable NALP1 protein signal in MonoMac6 cells expressing NALP1 shRNA (Fig. (Fig.2C).2C). These results corroborated the quantitative real-time PCR analysis and showed that NALP1 gene silencing was achieved effectively.
NALP1 knockdown reduces viability of human monocytic cells in vitro during infection with T. gondii.

To study if NALP1 gene knockdown affects the progression of T. gondii infection in human monocytic cells, the viability of in vitro-cultured wild-type or NALP1 or TetR shRNA-expressing MonoMac6 cells was analyzed at different time points of culture, using a colorimetric cell viability (WST-1 reagent-based) assay and a microscopic assay. The cell viability reagent WST-1 relies on the cleavage of tetrazolium salts to formazan by cellular mitochondrial dehydrogenases in the sample. The augmentation in dehydrogenase activity leads to an increase in the amount of formazan dye formed, which directly correlates with the number of metabolically active cells in the culture (3). The absorbances obtained from the T. gondii-infected MonoMac6 cells were normalized using the absorbances obtained from the respective uninfected MonoMac6 cells at each time point, and these data are presented as relative cell viability values (infected A420/uninfected A420). At day 0 (immediately after establishing the cultures), there were no significant differences in cell viability among the infected wild-type, NALP1 knockdown, and TetR shRNA-expressing cells (Fig. (Fig.3A).3A). Interestingly, after 1 day of culture, while the relative viability of the infected wildtype and TetR shRNA-expressing MonoMac6 cells had increased, the relative viability of the infected NALP1 knockdown cells was reduced significantly. On subsequent days of culture, this reduction in viability of the infected NALP1 knockdown cells was more pronounced, with almost no cell viability being observed by the fourth day of culture. In contrast, the numbers of infected wild-type and TetR shRNA-expressing cells started to decrease from day 2 of culture, but the relative cell viability was sustainably and significantly higher than that of the NALP1 knockdown cells at all subsequent days (Fig. (Fig.3A).3A). The values for the assay performed on wells containing parasites only (without MonoMac6 cells) were insignificant and declined rapidly over

time (data not shown), suggesting that in the infected cultures, almost all of the dehydrogenase enzyme activity was from MonoMac6 cells.

FIG. 3. Effect of NALP1 knockdown on viability of human monocytic cells during infection with T. gondii (A), rate of host cell infection (B), and rate of parasite multiplication within a parasitophorous vacuole in an infected cell (C). Wild-type MonoMac6 cells ... To visualize the state of the cells in culture and to assess how the relative cell viability values correlated with the state of the cells in culture, wild-type, NALP1 knockdown, and TetR shRNAexpressing MonoMac6 cells were cultured with or without parasites under the same conditions as those described above, except that the culture volumes were increased to 1 ml in 24-well plates. Cytospin Giemsa-stained smears made from these cultures at different time points showed morphologically intact cells with similar amounts of cells (in all categories of MonoMac6 cells) for both the infected and uninfected cells at day 0 of culture (Fig. (Fig.4).4). At 1 day postculture, infected NALP1 knockdown cells appeared sparse, while the rest of the cell categories looked about the same (with approximately the same density). By day 2 postculture, infected NALP1 knockdown cells had become sparser and were disintegrating. On the other hand, despite the presence of intracellular parasites in some of the infected wild-type and TetR shRNA-expressing cells, the cells looked relatively denser and morphologically normal, in contrast to the infected NALP1 knockdown cells. By day 3 postculture, the infected NALP1 knockdown cell cytospin preparations contained mostly cell debris, and by day 4, there were almost no observable morphologically intact MonoMac6 cells and only a small number of sparsely distributed extracellular parasites.

FIG. 4.

Giemsa-stained preparations of monocytic cells examined by light microscopy to determine the effect of NALP1 gene knockdown on monocytic cell viability when cells were cultured with or without parasites. Preparations of monocytic cells genetically modified ... Interestingly, while there was a gradual increase in the number of infected cells in the preparations of the wild-type and TetR shRNA-expressing cells at subsequent days, the cells remained mostly intact and dense, except for days 3 and 4, when the cell intensities decreased notably but were still denser than that of the infected NALP1 knockdown culture (Fig. (Fig.4).4). The uninfected cell cultures in all categories remained dense and morphologically intact at subsequent days of culture, except at day 4, when they were observed to be less dense, though the majority of the cells still appeared morphologically intact. These microscopic observations of cell culture progression provided visual evidence indicating that the decrease in cell viability (observed using the WST-1 assay) correlated with MonoMac6 cell death and that the results of both assays were consistent. Also, as can be seen in the representative images shown in Fig. Fig.44 (from one of three independent replicate studies), there were more parasites in the NALP1 knockdown cells and cultures than in the NALP1-sufficient cells. To assess the effect of NALP1 knockdown on the rate of progression of infection, the number of infected cells as well as the number of parasites in one parasitophorous vacuole of an infected cell at each time point was determined. Starting from day 1 postinfection, the rate of MonoMac6 cell infection with NALP1 knockdown was found to be significantly higher than those for both the wild-type and TetR shRNA-expressing cells (Fig. (Fig.3B).3B). Furthermore, infected MonoMac6 cells with NALP1 knockdown were observed to contain a larger number of parasites per parasitophorous vacuole than the NALP1-sufficient cells (Fig. (Fig.3C3C).
NALP1 knockdown modulates T. gondii-induced proinflammatory cytokine expression in MonoMac6 cells.

To assess the effect of NALP1 knockdown on the expression of proinflammatory cytokines in human monocytic cells during infection with T. gondii, wild-type MonoMac6 cells and MonoMac6 cells engineered to express either NALP1 gene shRNA or tetracycline repressor gene shRNA were cultured with or without parasites at an MOI of 1:4 (parasite/MonoMac6 cell ratio) for 36 h. The cells were harvested and total RNA extracted and reverse transcribed to synthesize first-strand cDNA. Quantitative real-time PCR was performed on each sample of the cDNA to

determine the transcript levels of the IL-1, IL-18, TNF-, IL-12, human actin, and GAPDH genes. The transcript values obtained for IL-1, IL-18, TNF-, IFN-, and IL-12 were normalized interchangeably with either human GAPDH or human actin transcript levels. When GAPDH was used for normalization, for all infected wild-type and TetR shRNA-expressing MonoMac6 cells, IL-1, IL-18, TNF-, and IL-12 were found to be significantly higher than the levels in noninfected cells (Fig. (Fig.55 A to D). The results were found to be consistent when human actin was used for normalization (see Fig. S1 in the supplemental material).

FIG. 5. Analysis of effects of NALP1 gene knockdown on expression of proinflammatory cytokines (IL-1, IL-18, TNF-, and IL-12) in a human monocytic cell line (MonoMac6). Quantitative real-time PCR was performed on cDNAs synthesized using equal ... Interestingly, in both infected and uninfected MonoMac6 cells with NALP1 knockdown, IL-1 expression was found to be lower than that in the uninfected wild-type or TetR shRNAexpressing MonoMac6 cells (Fig. (Fig.5A).5A). Moreover, there was no notable significant difference in IL-1 transcript levels between the infected and uninfected NALP1 knockdown MonoMac6 cells (Fig. (Fig.5A),5A), indicating that knockdown of NALP1 by itself dampens IL-1 transcription. In the case of IL-18 and IL-12 expression, while the transcript levels in the infected cells were about the same as those in the uninfected NALP1 knockdown cells, there was no observable dampening of the background expression (Fig. 5B and D). Specifically, IL-18 and IL12 transcript levels in the uninfected NALP1 knockdown cells were similar to those in the uninfected wild-type and TetR shRNA-expressing cells (Fig. (Fig.5B).5B). The expression of TNF was found to be consistently higher for all infected cell categories than for the noninfected cells (Fig. (Fig.5C),5C), although transcript levels in the infected NALP1 knockdown cells were slightly lower (though the difference was not statistically significant) than those in the infected wild-type and TetR shRNA-expressing cells. This suggested that the activation of TNF- expression during T. gondii infection was also affected, to some extent, by the lack of NALP1 inflammasome

activity. For all cell categories, transcripts of IFN- could not be detected (data not shown), which was expected because the cell cultures consisted of monocytic cells only, with no T cells. To ascertain that the observed low cytokine transcript levels in infected NALP1 knockdown cells were not simply a consequence of a global reduction in transcription due to reduced cell viability, the transcript levels of the BCL2A1 gene, whose transcription has been shown to be increased greatly by T. gondii infection in mammalian cells (12), were measured. For both the NALP1 knockdown and NALP1-sufficient MonoMac6 cells, BCL2A1 transcripts were found to be significantly higher in infected cells than in uninfected cells (Fig. (Fig.6A).6A). This result indicated that the reduction in cytokine expression was a result of NALP1 knockdown and not necessarily because of a global reduction in transcription.

FIG. 6. Analysis of effects of NALP1 gene knockdown on expression of caspase-1 and BCL2A1 in MonoMac6 cells infected with T. gondii. Quantitative real-time PCR was performed on cDNAs synthesized from equal amounts of total RNA from wild-type MonoMac6 cells (MC) ...
IL-1 protein accumulates as pro-IL-1 coding transcripts during NALP1 knockdown.

To determine the effect of NALP1 knockdown on the expression of mature IL-1, T. gondiiinfected and uninfected wild-type, TetR shRNA-expressing, and NALP1 knockdown monocytic cells were cultured for 36 h. The cells were harvested, and Western blotting was performed on the total cell lysates. While mature IL-1 was detectable as a strong signal of approximately 17 kDa in the T. gondii-infected wild-type and TetR shRNA-expressing cell lysates, it was barely detectable in the infected NALP1 knockdown cell lysates. Interestingly, in these lysates, in addition to the weak signal of mature IL-1 (17 kDa), there was a signal at 35 kDa consistent with pro-IL-1 (Fig. (Fig.77).

FIG. 7.

Western blot of the effect of NALP1 knockdown on the expression of IL-1 in MonoMac6 cells with or without T. gondii infection for 36 h. Wt, wild-type MonoMac6 cells; Tet, MonoMac6 cells engineered to express tetracycline repressor gene shRNA; ... To assess the direct role of NALP1 knockdown on the processing of pro-IL-1, the transcript levels of caspase-1 (a proteolytic enzyme that is activated by the NALP1 inflammasome to cleave and activate the secretion of mature IL-1 and IL-18) were determined. While the transcript levels of caspase-1 were found to be increased significantly in infected cells expressing NALP1, those in infected NALP1 knockdown cells remained at essentially the same level as in the uninfected cells (Fig. (Fig.6B).6B). These results indicated that the lack of processing of pro-IL-1 to mature IL-1 in NALP1 knockdown cells was because of the lack of caspase-1 activation by the NALP1 inflammasome. Go to:

DISCUSSION
Mammalian monocytes, a pleomorphic and pleiotropic population of circulating mononuclear cells, contribute to microbial defense by supplying tissues with macrophages and dendritic cell precursors (25, 33). T. gondii is an obligate intracellular parasite that resides in a vacuole in many different nucleated cell populations. Macrophage recruitment has been shown to be essential for initial restriction of T. gondii growth in murine models of toxoplasmosis and in human monocytes (2, 38). This restriction of parasites in murine macrophages is mediated through MyD88 signaling and through induction of IL-12 and IFN- (24, 39). NALP1 has been shown to be important in susceptibility to a variety of intracellular pathogens invading mononuclear phagocytic cells (8,36). In the present study, we found that NALP1 has alleles associated with susceptibility to congenital toxoplasmosis. To facilitate the elucidation of the role played by NALP1 and/or the NALP1 inflammasome during infection with T. gondii, we successfully genetically engineered human monocytic cells for NALP1 gene knockdown by RNAi. Our studies to determine the outcome of T. gondii infection in monocytic cells showed that NALP1 knockdown attenuated the progression of the infection, with accelerated loss of cell viability and eventual disintegration of monocytic cells. Consistent with these findings, T. gondii infection of monocytic cells with NALP1 knockdown could not induce upregulation of the

proinflammatory cytokines IL-1, IL-18, and IL-12. Intracellular accumulation of pro-IL-1 has been reported to be regulated transcriptionally, presumably via NF-B and mitogen-activated protein (MAP) kinase signaling (15, 16). An initial pathogen stimulus generated via PRRs causes accumulation of the intracellular stores of pro-IL-1 (5). The preceding event involves the intracellular pathogen activation of inflammasome assembly, leading to the activation of the cysteinyl aspartate protease caspase-1 and the cleavage of proforms of IL-1 and IL-18, enabling the release of active mature IL-1 and IL-18 (27). In our study, we observed upregulated expression of IL-1 transcripts in T. gondii-infected monocytic cells expressing NALP1, with a concomitant increase in the amount of mature IL-1 protein, consistent with reports that intracellular pathogens activate inflammasomes (5). On the other hand, our T. gondii-infected monocytic cells with NALP1 knockdown did not show upregulation of IL-1 transcript levels, nor did they show an increase in the amount of mature IL1 produced, suggesting that the NALP1 inflammasome plays a significant role in the maturation of IL-1 under these circumstances. Interestingly, we found that T. gondii infection did not lead to increased transcription of caspase-1 in NALP1 knockdown cells, implying that the NALP1 inflammasome plays a critical role in the processing and activation of caspase-1, which in turn leads to maturation of IL-1 and IL-18. Upon production, mature IL-1 activates its cognate receptor, the IL-1 receptor, which signals through the MyD88 adaptor leading to the initiation of the NF-B and MAP kinase pathways, with the ultimate expression of inflammatory cytokines that then fight infection (4, 9, 10). Our findings are in line with this report, in that the lack of mature IL-1 production in infected NALP1 knockdown cells was also associated with decreased expression of IL-12, IL-18, and TNF- (important inflammatory cytokines in combating this infection), indicating that the initiation of NF-B and MAP kinase pathways via IL-1 is crucial during T. gondiiinfection. While the lack of an increase in mature IL-1 was expected to be due to the absence of NALP1 inflammasome activation because of NALP1 knockdown, the absence of IL-1 transcript upregulation was striking and unexpected based on what is known about NALP1 in other systems to date. To protect against T. gondii infection, as stated above, an initial T. gondii stimulus generated via PRRs would cause an accumulation of intracellular stores of pro-IL-1 transcripts that would then be translated into pro-IL-1 protein. Because of the lack of NALP1

inflammasome activation of caspase-1, which in turn led to a lack of processing and secretion of mature IL-1 (because of NALP1 knockdown), pro-IL-1 accumulated, and this might have exerted negative feedback on the transcription of pro-IL-1. This could explain the low level of pro-IL-1 transcripts in infected NALP1 knockdown cells. This diminished transcription was not due to a global cellular reduction in transcription, since equal amounts of total RNA were used to synthesize cDNA for each cell type (wild type and knockdown). Furthermore, analysis of the transcript levels of the BCL2A1 gene, whose transcription has been shown to be increased greatly by T. gondii infection in mammalian cells (12), showed equal augmentation of BCL2A1 transcription in both NALP1 knockdown and NALP1-sufficient cells, implying that the reduction in cytokine expression was a result of NALP1 knockdown, not a global reduction in transcription. Transcripts for IL-18 were also found not to be upregulated in infected NALP1 knockdown cells. Just like that of pro-IL-1, pro-IL-18 transcription can be induced via NF-B, mediated by signaling through Toll-like receptors (TLRs) (22). TLR signaling then primes the assembly of caspase-1-containing inflammasomes that process the proforms of IL-1 and IL-18, two crucial proinflammatory cytokines (9). It can be postulated that in a manner similar to that for pro-IL1, the reduced inflammasome activity due to NALP1 knockdown leads to an accumulation of pro-IL-18 protein that may impart negative-feedback inhibition on the transcription of IL-18. It is noteworthy that, to date, three NLRs, namely, NALP1, NALP3, and Apaf, have been reported to form caspase-1-activating inflammasomes and therefore to be involved in the processing of the proforms of IL-1 and IL-18 (20, 22). Our finding that the knockdown of NALP1 significantly reduces the levels of mature IL-1 protein and the transcripts of both proforms of IL-1 and IL-18 suggests that the NALP1 inflammasome is a major player in the processing of these cytokines during T. gondiiinfection. We found that while infected monocytic cells with NALP1 knockdown had an accelerated loss of cell viability and cell disintegration, the infected wild-type cells had a gradual progression of parasite growth and host cell death, suggesting that the NALP1 inflammasome is critical for mediating the inhibition and killing of T. gondii in human monocytic cells. As is the case with other intracellular pathogens, T. gondii has been shown to prominently modulate host cell apoptosis, which may be critical for the course of infection. T. gondii has been reported to exert opposite effects on the host cell death program by either triggering (14, 29) or inhibiting (30)

apoptosis. Based on our findings and the available literature, it can be postulated that activation of the NALP1 inflammasome triggers a cascade of innate immune mechanisms, including production of proinflammatory cytokines and mechanisms leading to death of infected cells. Such mechanisms could include activation of pyroptosis and apoptosis. This would thereby limit parasite proliferation and infection of other cells. This would not take place in infected monocytic cells with NALP1 knockdown, resulting in the uncontrolled proliferation of parasites and, ultimately, in more rapid death of host cells than that of wild-type cells. The proinflammatory response could thus be beneficial. Alternatively, if the proinflammatory response is too robust, it might be harmful to the host. Such harm does not necessarily correlate with a high parasite burden. Underlying the success of T. gondii infection is a delicate balance between the host's immune response and the parasite's modulation of many of the intricate mechanisms that the host uses to try to kill the parasite, resulting in the survival of both host and parasite. Initial infection with T. gondii provokes the innate immune mechanism of the host, leading to production of substantial amounts of IL-12, TNF-, and IFN- (1). These cytokines play a crucial role in resistance to T. gondii, with ultimate generation of a robust Th1-biased CD4+ and CD8+ cell-mediated immune response which is also characterized by high levels of IL-12 and IFN- (31). In our study, unlike the infected wild-type cells, infected monocytic cells with NALP1 knockdown did not have upregulated expression of IL-12, suggesting that signaling of inflammatory cytokine expression mediated through the IL-1 receptor contributes significantly to the upregulation of IL-12 expression. Our findings are consistent with the NALP1 inflammasome playing a crucial role in mediating the expression during T. gondii infection of proinflammatory cytokines that are important in controlling pathogenesis in this infection. Understanding the mechanisms of inflammasome regulation and how their downstream pathways contribute to pathogenesis of T. gondii will be important for elucidating the host innate immune responses to T. gondii and the susceptibility to disease caused by this parasite.